Showing 1–50 of 2,047 results for author: Lin, D

Study of $\bar{K}^*(892)^0 η$ and $K_S^0 a_0(980)^0$ in the $D^{0} \to K_{S}^{0}π^0η$ decay

Authors: BESIII Collaboration, M. Ablikim, M. N. Achasov, P. Adlarson, X. C. Ai, R. Aliberti, A. Amoroso, Q. An, Y. Bai, O. Bakina, Y. Ban, H. -R. Bao, V. Batozskaya, K. Begzsuren, N. Berger, M. Berlowski, M. Bertani, D. Bettoni, F. Bianchi, E. Bianco, A. Bortone, I. Boyko, R. A. Briere, A. Brueggemann, H. Cai , et al. (658 additional authors not shown)

Abstract: We perform an amplitude analysis of the decay $D^0 \to K_S^0 π^0 η$ and measure its absolute branching fraction to be $(1.016 \pm 0.013_{\text {stat.}} \pm 0.014_{\text {syst.}})\%$. The analysis utilizes $20.3~\mathrm{fb}^{-1}$ of $e^{+}e^{-}$ collision data collected at a center-of-mass energy of 3.773~GeV with the BESIII detector. The branching fraction of the intermediate process… ▽ More We perform an amplitude analysis of the decay $D^0 \to K_S^0 π^0 η$ and measure its absolute branching fraction to be $(1.016 \pm 0.013_{\text {stat.}} \pm 0.014_{\text {syst.}})\%$. The analysis utilizes $20.3~\mathrm{fb}^{-1}$ of $e^{+}e^{-}$ collision data collected at a center-of-mass energy of 3.773~GeV with the BESIII detector. The branching fraction of the intermediate process $D^0 \rightarrow \bar{K}^*(892)^0 η$ is determined to be $(0.73 \pm 0.05_{\text {stat.}} \pm 0.03_{\text {syst.}}) \%$, which is $5 σ$ smaller than the result measured in $D^0 \to K^- π^+ η$ by the Belle experiment. As a result, the magnitudes of the $W$-exchange and QCD-penguin exchange amplitudes are found to be less than half of their current estimations. Furthermore, we determine $\mathcal{B}(D^0\to K_S^0a_0(980)^0, a_0(980)^0\to π^0η) = (9.88\pm 0.37_{\rm stat.}\pm 0.42_{\rm syst.})\times10^{-3}$, with a precision improved by a factor of 4.5 compared to the world average. △ Less

Submitted 29 December, 2025; originally announced December 2025.

SpidR-Adapt: A Universal Speech Representation Model for Few-Shot Adaptation

Authors: Mahi Luthra, Jiayi Shen, Maxime Poli, Angelo Ortiz, Yosuke Higuchi, Youssef Benchekroun, Martin Gleize, Charles-Eric Saint-James, Dongyan Lin, Phillip Rust, Angel Villar, Surya Parimi, Vanessa Stark, Rashel Moritz, Juan Pino, Yann LeCun, Emmanuel Dupoux

Abstract: Human infants, with only a few hundred hours of speech exposure, acquire basic units of new languages, highlighting a striking efficiency gap compared to the data-hungry self-supervised speech models. To address this gap, this paper introduces SpidR-Adapt for rapid adaptation to new languages using minimal unlabeled data. We cast such low-resource speech representation learning as a meta-learning… ▽ More Human infants, with only a few hundred hours of speech exposure, acquire basic units of new languages, highlighting a striking efficiency gap compared to the data-hungry self-supervised speech models. To address this gap, this paper introduces SpidR-Adapt for rapid adaptation to new languages using minimal unlabeled data. We cast such low-resource speech representation learning as a meta-learning problem and construct a multi-task adaptive pre-training (MAdaPT) protocol which formulates the adaptation process as a bi-level optimization framework. To enable scalable meta-training under this framework, we propose a novel heuristic solution, first-order bi-level optimization (FOBLO), avoiding heavy computation costs. Finally, we stabilize meta-training by using a robust initialization through interleaved supervision which alternates self-supervised and supervised objectives. Empirically, SpidR-Adapt achieves rapid gains in phonemic discriminability (ABX) and spoken language modeling (sWUGGY, sBLIMP, tSC), improving over in-domain language models after training on less than 1h of target-language audio, over $100\times$ more data-efficient than standard training. These findings highlight a practical, architecture-agnostic path toward biologically inspired, data-efficient representations. We open-source the training code and model checkpoints at https://github.com/facebookresearch/spidr-adapt. △ Less

Submitted 24 December, 2025; originally announced December 2025.

Structure-Preserving Nonlinear Sufficient Dimension Reduction for Tensors

Authors: Dianjun Lin, Bing Li, Lingzhou Xue

Abstract: We introduce two nonlinear sufficient dimension reduction methods for regressions with tensor-valued predictors. Our goal is two-fold: the first is to preserve the tensor structure when performing dimension reduction, particularly the meaning of the tensor modes, for improved interpretation; the second is to substantially reduce the number of parameters in dimension reduction, thereby achieving mo… ▽ More We introduce two nonlinear sufficient dimension reduction methods for regressions with tensor-valued predictors. Our goal is two-fold: the first is to preserve the tensor structure when performing dimension reduction, particularly the meaning of the tensor modes, for improved interpretation; the second is to substantially reduce the number of parameters in dimension reduction, thereby achieving model parsimony and enhancing estimation accuracy. Our two tensor dimension reduction methods echo the two commonly used tensor decomposition mechanisms: one is the Tucker decomposition, which reduces a larger tensor to a smaller one; the other is the CP-decomposition, which represents an arbitrary tensor as a sequence of rank-one tensors. We developed the Fisher consistency of our methods at the population level and established their consistency and convergence rates. Both methods are easy to implement numerically: the Tucker-form can be implemented through a sequence of least-squares steps, and the CP-form can be implemented through a sequence of singular value decompositions. We investigated the finite-sample performance of our methods and showed substantial improvement in accuracy over existing methods in simulations and two data applications. △ Less

Submitted 23 December, 2025; originally announced December 2025.

Comments: 34 pages

MSC Class: 62H25; 62G08

Observation and branching fraction measurements of $χ_{cJ}\to p \bar p K^0_S K^0_S$

Authors: BESIII Collaboration, M. Ablikim, M. N. Achasov, P. Adlarson, X. C. Ai, C. S. Akondi, R. Aliberti, A. Amoroso, Q. An, Y. H. An, Y. Bai, O. Bakina, Y. Ban, H. -R. Bao, X. L. Bao, V. Batozskaya, K. Begzsuren, N. Berger, M. Berlowski, M. B. Bertani, D. Bettoni, F. Bianchi, E. Bianco, A. Bortone, I. Boyko , et al. (705 additional authors not shown)

Abstract: Based on $(2.712\pm0.014)\times10^9$ $ψ(3686)$ events collected with the BESIII detector, the decays $χ_{cJ} \to p \bar{p} K^{0}_{S} K^{0}_{S}$ ($J=0,1,2$) are observed for the first time with statistical significances exceeding $5σ$.The measured branching fractions are $\mathcal{B}(χ_{c0}\to p \bar p K^{0}_{S} K^{0}_{S})=(6.94\pm0.30\pm0.38)\times10^{-5}$,… ▽ More Based on $(2.712\pm0.014)\times10^9$ $ψ(3686)$ events collected with the BESIII detector, the decays $χ_{cJ} \to p \bar{p} K^{0}_{S} K^{0}_{S}$ ($J=0,1,2$) are observed for the first time with statistical significances exceeding $5σ$.The measured branching fractions are $\mathcal{B}(χ_{c0}\to p \bar p K^{0}_{S} K^{0}_{S})=(6.94\pm0.30\pm0.38)\times10^{-5}$, $\mathcal{B}(χ_{c1}\to p \bar p K^{0}_{S} K^{0}_{S})=(1.30\pm0.12\pm0.08)\times10^{-5}$, and $\mathcal{B}(χ_{c2}\to p \bar p K^{0}_{S} K^{0}_{S})=(2.54\pm0.17\pm0.15)\times10^{-5}$, where the first uncertainties are statistical and the second systematic. In addition, no significant structures are observed in the $p\bar p$, $p K^0_S$, and $\bar p K^0_S$ mass spectra. △ Less

Submitted 29 December, 2025; v1 submitted 22 December, 2025; originally announced December 2025.

Comments: 10 Pages, 3 Figures, 3 Tables, Isospin discusstion updated

Report number: BAM-00840

Enhanced Information Security via Wave-Field Selectivity and Structured Wavefront Manipulation

Authors: Yufei Zhao, Deyu Lin, Qian Zhang, Haoyang Shi, Hong Niu, Afkar Mohamed Ismail, Yong Liang Guan, Chau Yuen

Abstract: In this paper, we propose a novel secure wireless transmission architecture that enables the co-existence of spatial field modulation (SFM) and digital bandpass modulation (DBM), utilizing multi-mode vortex waves and programmable meta-surfaces (PMS). Distinct from conventional joint modulation schemes, our approach establishes two logically independent transmission channels--SFM and DBM--thereby e… ▽ More In this paper, we propose a novel secure wireless transmission architecture that enables the co-existence of spatial field modulation (SFM) and digital bandpass modulation (DBM), utilizing multi-mode vortex waves and programmable meta-surfaces (PMS). Distinct from conventional joint modulation schemes, our approach establishes two logically independent transmission channels--SFM and DBM--thereby eliminating the need for joint signal design or time synchronization. Specifically, the orthogonality of vortex wave modes is exploited to construct a high-capacity multi-mode DBM channel, in which each mode carries modulated symbols independently. As the composite waveform passes through the PMS, energy from different vortex modes is spatially focused onto distinct positions, dynamically determined by the PMS configuration. This spatial mapping forms a unique lookup table that encodes additional information in the electro-magnetic (EM) field distribution, effectively enabling a second, concurrent SFM channel. To enhance physical-layer security, the DBM channel transmits encrypted symbols transformed via dynamic symbol-domain mapping, while the corresponding mapping relations--or key information--are carried by the SFM channel. This lightweight dual-channel encryption strategy provides strong confidentiality without requiring complex joint decoding. To validate the feasibility of the proposed architecture, we design and implement a proof-of-concept prototype system, and conduct experimental demonstrations under real-world wireless communication conditions. The experimental results confirm the effectiveness of the co-existent DBM-SFM design in achieving reliable and secure transmission. The proposed architecture offers a scalable, low-complexity, and secure transmission solution for future IoT networks, especially in scenarios demanding both spectral efficiency and physical-layer confidentiality. △ Less

Submitted 11 December, 2025; originally announced December 2025.

The Prism Hypothesis: Harmonizing Semantic and Pixel Representations via Unified Autoencoding

Authors: Weichen Fan, Haiwen Diao, Quan Wang, Dahua Lin, Ziwei Liu

Abstract: Deep representations across modalities are inherently intertwined. In this paper, we systematically analyze the spectral characteristics of various semantic and pixel encoders. Interestingly, our study uncovers a highly inspiring and rarely explored correspondence between an encoder's feature spectrum and its functional role: semantic encoders primarily capture low-frequency components that encode… ▽ More Deep representations across modalities are inherently intertwined. In this paper, we systematically analyze the spectral characteristics of various semantic and pixel encoders. Interestingly, our study uncovers a highly inspiring and rarely explored correspondence between an encoder's feature spectrum and its functional role: semantic encoders primarily capture low-frequency components that encode abstract meaning, whereas pixel encoders additionally retain high-frequency information that conveys fine-grained detail. This heuristic finding offers a unifying perspective that ties encoder behavior to its underlying spectral structure. We define it as the Prism Hypothesis, where each data modality can be viewed as a projection of the natural world onto a shared feature spectrum, just like the prism. Building on this insight, we propose Unified Autoencoding (UAE), a model that harmonizes semantic structure and pixel details via an innovative frequency-band modulator, enabling their seamless coexistence. Extensive experiments on ImageNet and MS-COCO benchmarks validate that our UAE effectively unifies semantic abstraction and pixel-level fidelity into a single latent space with state-of-the-art performance. △ Less

Submitted 22 December, 2025; originally announced December 2025.

Comments: Code link: https://github.com/WeichenFan/UAE

Zero-shot Reconstruction of In-Scene Object Manipulation from Video

Authors: Dixuan Lin, Tianyou Wang, Zhuoyang Pan, Yufu Wang, Lingjie Liu, Kostas Daniilidis

Abstract: We build the first system to address the problem of reconstructing in-scene object manipulation from a monocular RGB video. It is challenging due to ill-posed scene reconstruction, ambiguous hand-object depth, and the need for physically plausible interactions. Existing methods operate in hand centric coordinates and ignore the scene, hindering metric accuracy and practical use. In our method, we… ▽ More We build the first system to address the problem of reconstructing in-scene object manipulation from a monocular RGB video. It is challenging due to ill-posed scene reconstruction, ambiguous hand-object depth, and the need for physically plausible interactions. Existing methods operate in hand centric coordinates and ignore the scene, hindering metric accuracy and practical use. In our method, we first use data-driven foundation models to initialize the core components, including the object mesh and poses, the scene point cloud, and the hand poses. We then apply a two-stage optimization that recovers a complete hand-object motion from grasping to interaction, which remains consistent with the scene information observed in the input video. △ Less

Submitted 22 December, 2025; originally announced December 2025.

Cross sections measurement of $e^+e^-\to Ξ(1530)^0\barΞ^0 + c.c.$ and search for $ψ(3770)\toΞ(1530)^0\barΞ^0 + c.c.$

Authors: BESIII Colaboration, :, M. Ablikim, M. N. Achasov, P. Adlarson, X. C. Ai, R. Aliberti, A. Amoroso, Q. An, Y. Bai, O. Bakina, Y. Ban, H. -R. Bao, V. Batozskaya, K. Begzsuren, N. Berger, M. Berlowski, M. B. Bertani, D. Bettoni, F. Bianchi, E. Bianco, A. Bortone, I. Boyko, R. A. Briere, A. Brueggemann , et al. (680 additional authors not shown)

Abstract: Using $e^+e^-$ collision data collected with the BESIII detector corresponding to an integrated luminosity of 44.2 fb$^{-1}$, we measure the Born cross sections for the process $e^+e^- \to Ξ(1530)^{0} \barΞ^{0} + c.c.$ at forty-eight center-of-mass energies between 3.51 and 4.95 GeV. The potential signal from non-$D\bar{D}$ decays for $ψ(3770)$, i.e. $ψ(3770)\to Ξ(1530)^{0} \barΞ^{0}+ c.c.$, is in… ▽ More Using $e^+e^-$ collision data collected with the BESIII detector corresponding to an integrated luminosity of 44.2 fb$^{-1}$, we measure the Born cross sections for the process $e^+e^- \to Ξ(1530)^{0} \barΞ^{0} + c.c.$ at forty-eight center-of-mass energies between 3.51 and 4.95 GeV. The potential signal from non-$D\bar{D}$ decays for $ψ(3770)$, i.e. $ψ(3770)\to Ξ(1530)^{0} \barΞ^{0}+ c.c.$, is investigated by fitting the dressed cross section, no obvious signal is found. The upper limit of $\mathcal{B}(ψ(3770) \to Ξ(1530)^{0} \barΞ^{0}$) at 90\% confidence level is given. △ Less

Submitted 19 December, 2025; originally announced December 2025.

Comments: 9 Pages, 6 figures

End-to-End Training for Autoregressive Video Diffusion via Self-Resampling

Authors: Yuwei Guo, Ceyuan Yang, Hao He, Yang Zhao, Meng Wei, Zhenheng Yang, Weilin Huang, Dahua Lin

Abstract: Autoregressive video diffusion models hold promise for world simulation but are vulnerable to exposure bias arising from the train-test mismatch. While recent works address this via post-training, they typically rely on a bidirectional teacher model or online discriminator. To achieve an end-to-end solution, we introduce Resampling Forcing, a teacher-free framework that enables training autoregres… ▽ More Autoregressive video diffusion models hold promise for world simulation but are vulnerable to exposure bias arising from the train-test mismatch. While recent works address this via post-training, they typically rely on a bidirectional teacher model or online discriminator. To achieve an end-to-end solution, we introduce Resampling Forcing, a teacher-free framework that enables training autoregressive video models from scratch and at scale. Central to our approach is a self-resampling scheme that simulates inference-time model errors on history frames during training. Conditioned on these degraded histories, a sparse causal mask enforces temporal causality while enabling parallel training with frame-level diffusion loss. To facilitate efficient long-horizon generation, we further introduce history routing, a parameter-free mechanism that dynamically retrieves the top-k most relevant history frames for each query. Experiments demonstrate that our approach achieves performance comparable to distillation-based baselines while exhibiting superior temporal consistency on longer videos owing to native-length training. △ Less

Submitted 17 December, 2025; originally announced December 2025.

Comments: Project Page: https://guoyww.github.io/projects/resampling-forcing/

Measurements of the Absolute Branching Fraction of the Semileptonic Decay $\mathbf{Ξ^{-}\rightarrow Λe^- \barν_{e}}$ and the Axial Charge of the $\mathbfΞ^{-}$

Authors: BESIII Collaboration, M. Ablikim, M. N. Achasov, P. Adlarson, X. C. Ai, R. Aliberti, A. Amoroso, Q. An, Y. Bai, O. Bakina, Y. Ban, H. -R. Bao, V. Batozskaya, K. Begzsuren, N. Berger, M. Berlowski, M. B. Bertani, D. Bettoni, F. Bianchi, E. Bianco, A. Bortone, I. Boyko, R. A. Briere, A. Brueggemann, H. Cai , et al. (683 additional authors not shown)

Abstract: Using $(10087\pm44)\times10^6$ $J/ψ$ events collected with the BESIII detector, we study the semileptonic decay $Ξ^{-}\rightarrow Λe^- \barν_{e}$ for the first time at an electron-positron collider. The absolute branching fraction is determined for the first time to be $(3.60\pm0.40_{\mathrm{stat}}\pm0.10_{\mathrm{syst}})\times10^{-4}$, which is 3.9 standard deviations below the world average. In… ▽ More Using $(10087\pm44)\times10^6$ $J/ψ$ events collected with the BESIII detector, we study the semileptonic decay $Ξ^{-}\rightarrow Λe^- \barν_{e}$ for the first time at an electron-positron collider. The absolute branching fraction is determined for the first time to be $(3.60\pm0.40_{\mathrm{stat}}\pm0.10_{\mathrm{syst}})\times10^{-4}$, which is 3.9 standard deviations below the world average. In addition, using an 11-dimensional angular analysis, the axial-vector to vector coupling $g_{av}$ is determined to be $0.18\pm0.07_{\mathrm{stat}}\pm0.02_{\mathrm{syst}}$. These results are used to test various SU(3)-flavour effective models. Under the SU(3)-flavour symmetry limit, the axial charge is found to be $g_A^H = 0.22\pm0.08_{\mathrm{stat}}\pm0.02_{\mathrm{syst}}$. Despite using only 5\% of the statistics of previous experiments, this analysis achieves a comparable precision for the axial charge. △ Less

Submitted 17 December, 2025; originally announced December 2025.

Comments: 10 pages, 4 figures

Search for the decays $X(3872)\to K_{S}^{0}K^{\pm}π^{\mp}$ and $K^*(892)\bar{K}$ at BESIII

Authors: BESIII Collaboration, M. Ablikim, M. N. Achasov, P. Adlarson, X. C. Ai, R. Aliberti, A. Amoroso, Q. An, Y. Bai, O. Bakina, Y. Ban, H. -R. Bao, X. L. Bao, V. Batozskaya, K. Begzsuren, N. Berger, M. Berlowski, M. B. Bertani, D. Bettoni, F. Bianchi, E. Bianco, A. Bortone, I. Boyko, R. A. Briere, A. Brueggemann , et al. (684 additional authors not shown)

Abstract: Using a 10.9 fb$^{-1}$ data sample collected by the BESIII detector at center-of-mass energies from 4.16 to 4.34 GeV, we search for the charmless decays $X(3872) \to K_{S}^{0}K^{\pm}π^{\mp}$ and $K^*(892)\bar{K}$, where the $X(3872)$ is produced via the radiative process $e^+e^- \to γX(3872)$. No significant signal is observed. We set upper limits on the relative branching fractions… ▽ More Using a 10.9 fb$^{-1}$ data sample collected by the BESIII detector at center-of-mass energies from 4.16 to 4.34 GeV, we search for the charmless decays $X(3872) \to K_{S}^{0}K^{\pm}π^{\mp}$ and $K^*(892)\bar{K}$, where the $X(3872)$ is produced via the radiative process $e^+e^- \to γX(3872)$. No significant signal is observed. We set upper limits on the relative branching fractions $\mathcal{B}[X(3872)\to K_S K^{\pm} π^{\mp}]/\mathcal{B}[X(3872)\toπ^+π^- J/ψ] <0.07$ and $\mathcal{B}[X(3872)\to K^* (892)\bar{K}]/\mathcal{B}[X(3872)\to π^+π^- J/ψ] <0.10$ at the 90$\%$ confidence level. Additionally, upper limits on the product of the cross section $σ[e^+e^-\toγX(3872)]$ and the branching fractions $\mathcal{B}[X(3872)\to K_{S}^{0}K^{\pm}π^{\mp}]$ and $\mathcal{B}[X(3872)\to K^*(892)\bar{K}]$ are reported at each energy point. In all cases, $K^*(892)\bar{K}$ refers to the sum of the modes $K^*(892)^+K^{-}+\text{c.c.}$ and $K^*(892)^0\bar{K}^0+\text{c.c.}$, where c.c. denotes the corresponding charge-conjugate modes. △ Less

Submitted 17 December, 2025; originally announced December 2025.

Comments: 11 pages, 1 figure

TalkVerse: Democratizing Minute-Long Audio-Driven Video Generation

Authors: Zhenzhi Wang, Jian Wang, Ke Ma, Dahua Lin, Bing Zhou

Abstract: We introduce TalkVerse, a large-scale, open corpus for single-person, audio-driven talking video generation designed to enable fair, reproducible comparison across methods. While current state-of-the-art systems rely on closed data or compute-heavy models, TalkVerse offers 2.3 million high-resolution (720p/1080p) audio-video synchronized clips totaling 6.3k hours. These are curated from over 60k h… ▽ More We introduce TalkVerse, a large-scale, open corpus for single-person, audio-driven talking video generation designed to enable fair, reproducible comparison across methods. While current state-of-the-art systems rely on closed data or compute-heavy models, TalkVerse offers 2.3 million high-resolution (720p/1080p) audio-video synchronized clips totaling 6.3k hours. These are curated from over 60k hours of video via a transparent pipeline that includes scene-cut detection, aesthetic assessment, strict audio-visual synchronization checks, and comprehensive annotations including 2D skeletons and structured visual/audio-style captions. Leveraging TalkVerse, we present a reproducible 5B DiT baseline built on Wan2.2-5B. By utilizing a video VAE with a high downsampling ratio and a sliding window mechanism with motion-frame context, our model achieves minute-long generation with low drift. It delivers comparable lip-sync and visual quality to the 14B Wan-S2V model but with 10$\times$ lower inference cost. To enhance storytelling in long videos, we integrate an MLLM director to rewrite prompts based on audio and visual cues. Furthermore, our model supports zero-shot video dubbing via controlled latent noise injection. We open-source the dataset, training recipes, and 5B checkpoints to lower barriers for research in audio-driven human video generation. Project Page: https://zhenzhiwang.github.io/talkverse/ △ Less

Submitted 16 December, 2025; originally announced December 2025.

Comments: open-sourced single-person full-body talking video generation dataset, training code and checkpoints

Measurements of the branching fractions of $χ_{cJ}\to φφη, φφη^{\prime}$ and $φK^+K^-η$

Authors: BESIII Collaboration, M. Ablikim, M. N. Achasov, P. Adlarson, X. C. Ai, R. Aliberti, A. Amoroso, Q. An, Y. Bai, O. Bakina, Y. Ban, H. -R. Bao, V. Batozskaya, K. Begzsuren, N. Berger, M. Berlowski, M. B. Bertani, D. Bettoni, F. Bianchi, E. Bianco, A. Bortone, I. Boyko, R. A. Briere, A. Brueggemann, H. Cai , et al. (681 additional authors not shown)

Abstract: Using a sample of $(2712.4 \pm 14.3)\times 10^6 ~ψ$(3686) events collected by the BESIII detector at the BEPCII collider, we measure the branching fractions of the decays $χ_{cJ}\to φφη,~φφη^{\prime}$, and~$φK^+K^-η$ ($J = 0, 1, 2$). The obtained branching fractions are $\mathcal{B}(χ_{c0} \to φφη) = (7.40 \pm 0.23 \pm 0.55)\times10^{-4}$,… ▽ More Using a sample of $(2712.4 \pm 14.3)\times 10^6 ~ψ$(3686) events collected by the BESIII detector at the BEPCII collider, we measure the branching fractions of the decays $χ_{cJ}\to φφη,~φφη^{\prime}$, and~$φK^+K^-η$ ($J = 0, 1, 2$). The obtained branching fractions are $\mathcal{B}(χ_{c0} \to φφη) = (7.40 \pm 0.23 \pm 0.55)\times10^{-4}$, $\mathcal{B}(χ_{c1} \to φφη) = (3.33 \pm 0.14 \pm 0.25)\times10^{-4}$, $\mathcal{B}(χ_{c2} \to φφη) = (5.46 \pm 0.17 \pm 0.40)\times10^{-4}$, $\mathcal{B}(χ_{c0} \to φφη^\prime) = (2.96 \pm 0.23 \pm 0.29)\times10^{-4}$, $\mathcal{B}(χ_{c1} \to φφη^\prime) = (0.69 \pm 0.10 \pm 0.08)\times10^{-4}$, $\mathcal{B}(χ_{c2} \to φφη^\prime) = (0.65 \pm 0.09 \pm 0.07)\times10^{-4}$, $\mathcal{B}(χ_{c0} \to φK^+K^-η) = (1.23 \pm 0.08 \pm 0.10)\times10^{-4}$, $\mathcal{B}(χ_{c1} \to φK^+K^-η) = (1.00 \pm 0.07 \pm 0.07)\times10^{-4}$, and $\mathcal{B}(χ_{c2} \to φK^+K^-η) = (1.82 \pm 0.09 \pm 0.14)\times10^{-4}$, where $K^+K^-$ is not from the decay of a $φ$ meson, the first uncertainties are statistical and the second systematic. The branching fractions of $χ_{cJ}\to φφη$ are measured with precision improved by factors of $1.5-1.9$, and those of $χ_{cJ}\to φφη^\prime$ and $φK^+K^-η$ are measured for the first time. △ Less

Submitted 16 December, 2025; originally announced December 2025.

Comments: 12 pages, 8 figures, 3 tables

Report number: BAM-00909

SS4D: Native 4D Generative Model via Structured Spacetime Latents

Authors: Zhibing Li, Mengchen Zhang, Tong Wu, Jing Tan, Jiaqi Wang, Dahua Lin

Abstract: We present SS4D, a native 4D generative model that synthesizes dynamic 3D objects directly from monocular video. Unlike prior approaches that construct 4D representations by optimizing over 3D or video generative models, we train a generator directly on 4D data, achieving high fidelity, temporal coherence, and structural consistency. At the core of our method is a compressed set of structured spac… ▽ More We present SS4D, a native 4D generative model that synthesizes dynamic 3D objects directly from monocular video. Unlike prior approaches that construct 4D representations by optimizing over 3D or video generative models, we train a generator directly on 4D data, achieving high fidelity, temporal coherence, and structural consistency. At the core of our method is a compressed set of structured spacetime latents. Specifically, (1) To address the scarcity of 4D training data, we build on a pre-trained single-image-to-3D model, preserving strong spatial consistency. (2) Temporal consistency is enforced by introducing dedicated temporal layers that reason across frames. (3) To support efficient training and inference over long video sequences, we compress the latent sequence along the temporal axis using factorized 4D convolutions and temporal downsampling blocks. In addition, we employ a carefully designed training strategy to enhance robustness against occlusion △ Less

Submitted 16 December, 2025; originally announced December 2025.

Comments: ToG(Siggraph Asia 2025)

Journal ref: ACM Transactions on Graphics, 44(6): Article 244, 12 pages, December 2025

OpenDataArena: A Fair and Open Arena for Benchmarking Post-Training Dataset Value

Authors: Mengzhang Cai, Xin Gao, Yu Li, Honglin Lin, Zheng Liu, Zhuoshi Pan, Qizhi Pei, Xiaoran Shang, Mengyuan Sun, Zinan Tang, Xiaoyang Wang, Zhanping Zhong, Yun Zhu, Dahua Lin, Conghui He, Lijun Wu

Abstract: The rapid evolution of Large Language Models (LLMs) is predicated on the quality and diversity of post-training datasets. However, a critical dichotomy persists: while models are rigorously benchmarked, the data fueling them remains a black box--characterized by opaque composition, uncertain provenance, and a lack of systematic evaluation. This opacity hinders reproducibility and obscures the caus… ▽ More The rapid evolution of Large Language Models (LLMs) is predicated on the quality and diversity of post-training datasets. However, a critical dichotomy persists: while models are rigorously benchmarked, the data fueling them remains a black box--characterized by opaque composition, uncertain provenance, and a lack of systematic evaluation. This opacity hinders reproducibility and obscures the causal link between data characteristics and model behaviors. To bridge this gap, we introduce OpenDataArena (ODA), a holistic and open platform designed to benchmark the intrinsic value of post-training data. ODA establishes a comprehensive ecosystem comprising four key pillars: (i) a unified training-evaluation pipeline that ensures fair, open comparisons across diverse models (e.g., Llama, Qwen) and domains; (ii) a multi-dimensional scoring framework that profiles data quality along tens of distinct axes; (iii) an interactive data lineage explorer to visualize dataset genealogy and dissect component sources; and (iv) a fully open-source toolkit for training, evaluation, and scoring to foster data research. Extensive experiments on ODA--covering over 120 training datasets across multiple domains on 22 benchmarks, validated by more than 600 training runs and 40 million processed data points--reveal non-trivial insights. Our analysis uncovers the inherent trade-offs between data complexity and task performance, identifies redundancy in popular benchmarks through lineage tracing, and maps the genealogical relationships across datasets. We release all results, tools, and configurations to democratize access to high-quality data evaluation. Rather than merely expanding a leaderboard, ODA envisions a shift from trial-and-error data curation to a principled science of Data-Centric AI, paving the way for rigorous studies on data mixing laws and the strategic composition of foundation models. △ Less

Submitted 15 December, 2025; originally announced December 2025.

DrivePI: Spatial-aware 4D MLLM for Unified Autonomous Driving Understanding, Perception, Prediction and Planning

Authors: Zhe Liu, Runhui Huang, Rui Yang, Siming Yan, Zining Wang, Lu Hou, Di Lin, Xiang Bai, Hengshuang Zhao

Abstract: Although multi-modal large language models (MLLMs) have shown strong capabilities across diverse domains, their application in generating fine-grained 3D perception and prediction outputs in autonomous driving remains underexplored. In this paper, we propose DrivePI, a novel spatial-aware 4D MLLM that serves as a unified Vision-Language-Action (VLA) framework that is also compatible with vision-ac… ▽ More Although multi-modal large language models (MLLMs) have shown strong capabilities across diverse domains, their application in generating fine-grained 3D perception and prediction outputs in autonomous driving remains underexplored. In this paper, we propose DrivePI, a novel spatial-aware 4D MLLM that serves as a unified Vision-Language-Action (VLA) framework that is also compatible with vision-action (VA) models. Our method jointly performs spatial understanding, 3D perception (i.e., 3D occupancy), prediction (i.e., occupancy flow), and planning (i.e., action outputs) in parallel through end-to-end optimization. To obtain both precise geometric information and rich visual appearance, our approach integrates point clouds, multi-view images, and language instructions within a unified MLLM architecture. We further develop a data engine to generate text-occupancy and text-flow QA pairs for 4D spatial understanding. Remarkably, with only a 0.5B Qwen2.5 model as MLLM backbone, DrivePI as a single unified model matches or exceeds both existing VLA models and specialized VA models. Specifically, compared to VLA models, DrivePI outperforms OpenDriveVLA-7B by 2.5% mean accuracy on nuScenes-QA and reduces collision rate by 70% over ORION (from 0.37% to 0.11%) on nuScenes. Against specialized VA models, DrivePI surpasses FB-OCC by 10.3 RayIoU for 3D occupancy on OpenOcc, reduces the mAVE from 0.591 to 0.509 for occupancy flow on OpenOcc, and achieves 32% lower L2 error than VAD (from 0.72m to 0.49m) for planning on nuScenes. Code will be available at https://github.com/happinesslz/DrivePI △ Less

Submitted 14 December, 2025; originally announced December 2025.

Amplitude Analysis and Branching Fraction Measurement of $D^+ \to π^+π^0π^0$

Authors: BESIII Collaboration, M. Ablikim, M. N. Achasov, P. Adlarson, X. C. Ai, R. Aliberti, A. Amoroso, Q. An, Y. Bai, O. Bakina, Y. Ban, H. -R. Bao, V. Batozskaya, K. Begzsuren, N. Berger, M. Berlowski, M. B. Bertani, D. Bettoni, F. Bianchi, E. Bianco, A. Bortone, I. Boyko, R. A. Briere, A. Brueggemann, H. Cai , et al. (684 additional authors not shown)

Abstract: We present the first amplitude analysis of the hadronic decay $D^+\toπ^+π^0π^0$, using $e^{+}e^{-}$ collision data collected with the BESIII detector at a center-of-mass energy of 3.773~GeV, corresponding to an integrated luminosity of 20.3~fb$^{-1}$. The fit fractions of the intermediate processes are measured, in which the $D^+ \to ρ(770)^+π^0$ component is found to be dominant with a branching… ▽ More We present the first amplitude analysis of the hadronic decay $D^+\toπ^+π^0π^0$, using $e^{+}e^{-}$ collision data collected with the BESIII detector at a center-of-mass energy of 3.773~GeV, corresponding to an integrated luminosity of 20.3~fb$^{-1}$. The fit fractions of the intermediate processes are measured, in which the $D^+ \to ρ(770)^+π^0$ component is found to be dominant with a branching fraction of $(3.08\kern0.15em\pm\kern0.15em0.10_{\rm stat.}\pm0.05_{\rm syst.})\times10^{-3}$. Based on the amplitude analysis, the branching fraction of $D^+ \to π^+π^0π^0$ is measured to be $(4.84\kern0.1em\pm\kern0.1em0.05_{\rm stat.}\kern0.1em\pm\kern0.1em0.05_{\rm syst.})\times10^{-3}$. In addition, the CP asymmetries, both for specific amplitudes and integrated over the entire phase space, are measured. △ Less

Submitted 13 December, 2025; originally announced December 2025.

MMSI-Video-Bench: A Holistic Benchmark for Video-Based Spatial Intelligence

Authors: Jingli Lin, Runsen Xu, Shaohao Zhu, Sihan Yang, Peizhou Cao, Yunlong Ran, Miao Hu, Chenming Zhu, Yiman Xie, Yilin Long, Wenbo Hu, Dahua Lin, Tai Wang, Jiangmiao Pang

Abstract: Spatial understanding over continuous visual input is crucial for MLLMs to evolve into general-purpose assistants in physical environments. Yet there is still no comprehensive benchmark that holistically assesses the progress toward this goal. In this work, we introduce MMSI-Video-Bench, a fully human-annotated benchmark for video-based spatial intelligence in MLLMs. It operationalizes a four-leve… ▽ More Spatial understanding over continuous visual input is crucial for MLLMs to evolve into general-purpose assistants in physical environments. Yet there is still no comprehensive benchmark that holistically assesses the progress toward this goal. In this work, we introduce MMSI-Video-Bench, a fully human-annotated benchmark for video-based spatial intelligence in MLLMs. It operationalizes a four-level framework, Perception, Planning, Prediction, and Cross-Video Reasoning, through 1,106 questions grounded in 1,278 clips from 25 datasets and in-house videos. Each item is carefully designed and reviewed by 3DV experts with explanatory rationales to ensure precise, unambiguous grounding. Leveraging its diverse data sources and holistic task coverage, MMSI-Video-Bench also supports three domain-oriented sub-benchmarks (Indoor Scene Perception Bench, Robot Bench and Grounding Bench) for targeted capability assessment. We evaluate 25 strong open-source and proprietary MLLMs, revealing a striking human--AI gap: many models perform near chance, and the best reasoning model lags humans by nearly 60%. We further find that spatially fine-tuned models still fail to generalize effectively on our benchmark. Fine-grained error analysis exposes systematic failures in geometric reasoning, motion grounding, long-horizon prediction, and cross-video correspondence. We also show that typical frame-sampling strategies transfer poorly to our reasoning-intensive benchmark, and that neither 3D spatial cues nor chain-of-thought prompting yields meaningful gains. We expect our benchmark to establish a solid testbed for advancing video-based spatial intelligence. △ Less

Submitted 11 December, 2025; originally announced December 2025.

OPV: Outcome-based Process Verifier for Efficient Long Chain-of-Thought Verification

Authors: Zijian Wu, Lingkai Kong, Wenwei Zhang, Songyang Gao, Yuzhe Gu, Zhongrui Cai, Tianyou Ma, Yuhong Liu, Zhi Wang, Runyuan Ma, Guangyu Wang, Wei Li, Conghui He, Dahua Lin, Kai Chen

Abstract: Large language models (LLMs) have achieved significant progress in solving complex reasoning tasks by Reinforcement Learning with Verifiable Rewards (RLVR). This advancement is also inseparable from the oversight automated by reliable verifiers. However, current outcome-based verifiers (OVs) are unable to inspect the unreliable intermediate steps in the long reasoning chains of thought (CoTs). Mea… ▽ More Large language models (LLMs) have achieved significant progress in solving complex reasoning tasks by Reinforcement Learning with Verifiable Rewards (RLVR). This advancement is also inseparable from the oversight automated by reliable verifiers. However, current outcome-based verifiers (OVs) are unable to inspect the unreliable intermediate steps in the long reasoning chains of thought (CoTs). Meanwhile, current process-based verifiers (PVs) have difficulties in reliably detecting errors in the complex long CoTs, limited by the scarcity of high-quality annotations due to the prohibitive costs of human annotations. Therefore, we propose the Outcome-based Process Verifier (OPV), which verifies the rationale process of summarized outcomes from long CoTs to achieve both accurate and efficient verification and enable large-scale annotation. To empower the proposed verifier, we adopt an iterative active learning framework with expert annotations to progressively improve the verification capability of OPV with fewer annotation costs. Specifically, in each iteration, the most uncertain cases of the current best OPV are annotated and then subsequently used to train a new OPV through Rejection Fine-Tuning (RFT) and RLVR for the next round. Extensive experiments demonstrate OPV's superior performance and broad applicability. It achieves new state-of-the-art results on our held-out OPV-Bench, outperforming much larger open-source models such as Qwen3-Max-Preview with an F1 score of 83.1 compared to 76.3. Furthermore, OPV effectively detects false positives within synthetic dataset, closely align with expert assessment. When collaborating with policy models, OPV consistently yields performance gains, e.g., raising the accuracy of DeepSeek-R1-Distill-Qwen-32B from 55.2% to 73.3% on AIME2025 as the compute budget scales. △ Less

Submitted 11 December, 2025; originally announced December 2025.

Long-horizon Reasoning Agent for Olympiad-Level Mathematical Problem Solving

Authors: Songyang Gao, Yuzhe Gu, Zijian Wu, Lingkai Kong, Wenwei Zhang, Zhongrui Cai, Fan Zheng, Tianyou Ma, Junhao Shen, Haiteng Zhao, Duanyang Zhang, Huilun Zhang, Kuikun Liu, Chengqi Lyu, Yanhui Duan, Chiyu Chen, Ningsheng Ma, Jianfei Gao, Han Lyu, Dahua Lin, Kai Chen

Abstract: Large Reasoning Models (LRMs) have expanded the mathematical reasoning frontier through Chain-of-Thought (CoT) techniques and Reinforcement Learning with Verifiable Rewards (RLVR), capable of solving AIME-level problems. However, the performance of LRMs is heavily dependent on the extended reasoning context length. For solving ultra-hard problems like those in the International Mathematical Olympi… ▽ More Large Reasoning Models (LRMs) have expanded the mathematical reasoning frontier through Chain-of-Thought (CoT) techniques and Reinforcement Learning with Verifiable Rewards (RLVR), capable of solving AIME-level problems. However, the performance of LRMs is heavily dependent on the extended reasoning context length. For solving ultra-hard problems like those in the International Mathematical Olympiad (IMO), the required reasoning complexity surpasses the space that an LRM can explore in a single round. Previous works attempt to extend the reasoning context of LRMs but remain prompt-based and built upon proprietary models, lacking systematic structures and training pipelines. Therefore, this paper introduces Intern-S1-MO, a long-horizon math agent that conducts multi-round hierarchical reasoning, composed of an LRM-based multi-agent system including reasoning, summary, and verification. By maintaining a compact memory in the form of lemmas, Intern-S1-MO can more freely explore the lemma-rich reasoning spaces in multiple reasoning stages, thereby breaking through the context constraints for IMO-level math problems. Furthermore, we propose OREAL-H, an RL framework for training the LRM using the online explored trajectories to simultaneously bootstrap the reasoning ability of LRM and elevate the overall performance of Intern-S1-MO. Experiments show that Intern-S1-MO can obtain 26 out of 35 points on the non-geometry problems of IMO2025, matching the performance of silver medalists. It also surpasses the current advanced LRMs on inference benchmarks such as HMMT2025, AIME2025, and CNMO2025. In addition, our agent officially participates in CMO2025 and achieves a score of 102/126 under the judgment of human experts, reaching the gold medal level. △ Less

Submitted 11 December, 2025; v1 submitted 11 December, 2025; originally announced December 2025.

Achieving Olympia-Level Geometry Large Language Model Agent via Complexity Boosting Reinforcement Learning

Authors: Haiteng Zhao, Junhao Shen, Yiming Zhang, Songyang Gao, Kuikun Liu, Tianyou Ma, Fan Zheng, Dahua Lin, Wenwei Zhang, Kai Chen

Abstract: Large language model (LLM) agents exhibit strong mathematical problem-solving abilities and can even solve International Mathematical Olympiad (IMO) level problems with the assistance of formal proof systems. However, due to weak heuristics for auxiliary constructions, AI for geometry problem solving remains dominated by expert models such as AlphaGeometry 2, which rely heavily on large-scale data… ▽ More Large language model (LLM) agents exhibit strong mathematical problem-solving abilities and can even solve International Mathematical Olympiad (IMO) level problems with the assistance of formal proof systems. However, due to weak heuristics for auxiliary constructions, AI for geometry problem solving remains dominated by expert models such as AlphaGeometry 2, which rely heavily on large-scale data synthesis and search for both training and evaluation. In this work, we make the first attempt to build a medalist-level LLM agent for geometry and present InternGeometry. InternGeometry overcomes the heuristic limitations in geometry by iteratively proposing propositions and auxiliary constructions, verifying them with a symbolic engine, and reflecting on the engine's feedback to guide subsequent proposals. A dynamic memory mechanism enables InternGeometry to conduct more than two hundred interactions with the symbolic engine per problem. To further accelerate learning, we introduce Complexity-Boosting Reinforcement Learning (CBRL), which gradually increases the complexity of synthesized problems across training stages. Built on InternThinker-32B, InternGeometry solves 44 of 50 IMO geometry problems (2000-2024), exceeding the average gold medalist score (40.9), using only 13K training examples, just 0.004% of the data used by AlphaGeometry 2, demonstrating the potential of LLM agents on expert-level geometry tasks. InternGeometry can also propose novel auxiliary constructions for IMO problems that do not appear in human solutions. We will release the model, data, and symbolic engine to support future research. △ Less

Submitted 12 December, 2025; v1 submitted 11 December, 2025; originally announced December 2025.

First measurement of the absolute branching fractions of $Σ^+$ nonleptonic decays and test of the $ΔI = 1/2$ rule % $Σ^+ \to p π^0$ and $Σ^+ \to n π^+$

Authors: BESIII Collaboration, M. Ablikim, M. N. Achasov, P. Adlarson, X. C. Ai, R. Aliberti, A. Amoroso, Q. An, Y. Bai, O. Bakina, Y. Ban, H. -R. Bao, X. L. Bao, V. Batozskaya, K. Begzsuren, N. Berger, M. Berlowski, M. B. Bertani, D. Bettoni, F. Bianchi, E. Bianco, A. Bortone, I. Boyko, R. A. Briere, A. Brueggemann , et al. (689 additional authors not shown)

Abstract: Based on $(10087 \pm 44) \times 10^6$ $J/ψ$ events collected by the BESIII detector at the center-of-mass energy $\sqrt{s} = 3.097$ GeV, the first absolute measurement of the branching fractions for the decays $Σ^+ \to p π^0$ and $Σ^+ \to n π^+$ is performed. The branching fractions are determined to be $B_{Σ^+ \to p π^0} = (49.79 \pm 0.06 \pm 0.22)\%$ and… ▽ More Based on $(10087 \pm 44) \times 10^6$ $J/ψ$ events collected by the BESIII detector at the center-of-mass energy $\sqrt{s} = 3.097$ GeV, the first absolute measurement of the branching fractions for the decays $Σ^+ \to p π^0$ and $Σ^+ \to n π^+$ is performed. The branching fractions are determined to be $B_{Σ^+ \to p π^0} = (49.79 \pm 0.06 \pm 0.22)\%$ and $B_{Σ^+ \to n π^+} = (49.87 \pm 0.05 \pm 0.29)\%$, where the first uncertainties are statistical and the second systematic. These results show significant deviations from the PDG values for both decays, with differences of 4.4$σ$ for $Σ^+ \to p π^0$ and 3.4$σ$ for $Σ^+ \to n π^+$. Furthermore, the $ΔI = 1/2$ rule is tested in nonleptonic $Σ^\pm$ decays. The observed results deviate from zero by more than $5σ$, indicating the presence of the $ΔI = 3/2$ transition amplitude in the $Σ$ hyperon decays. △ Less

Submitted 10 December, 2025; originally announced December 2025.

Measurement of the branching fraction of $η\to μ^+ μ^-$ and search for $η\to e^+ e^-$

Authors: BESIII Collaboration, M. Ablikim, M. N. Achasov, P. Adlarson, X. C. Ai, R. Aliberti, A. Amoroso, Q. An, Y. Bai, O. Bakina, Y. Ban, H. -R. Bao, V. Batozskaya, K. Begzsuren, N. Berger, M. Berlowski, M. Bertani, D. Bettoni, F. Bianchi, E. Bianco, A. Bortone, I. Boyko, R. A. Briere, A. Brueggemann, H. Cai , et al. (706 additional authors not shown)

Abstract: We analyze the decay of $η\rightarrow \ell^+\ell^-(\ell=e, μ)$ via $J/ψ\rightarrowγη'$ and $η'\rightarrowπ^+π^-η$, based on (10087 $\pm$ 44) $\times$ 10$^{6}$ $J/ψ$ events collected with the BESIII detector at the BEPCII storage rings. The branching fraction of $η\rightarrowμ^+ μ^-$ is measured to be $(5.8 \pm 1.0_{\rm stat} \pm 0.2_{\rm syst}) \times 10^{-6}$, which is consistent with the previou… ▽ More We analyze the decay of $η\rightarrow \ell^+\ell^-(\ell=e, μ)$ via $J/ψ\rightarrowγη'$ and $η'\rightarrowπ^+π^-η$, based on (10087 $\pm$ 44) $\times$ 10$^{6}$ $J/ψ$ events collected with the BESIII detector at the BEPCII storage rings. The branching fraction of $η\rightarrowμ^+ μ^-$ is measured to be $(5.8 \pm 1.0_{\rm stat} \pm 0.2_{\rm syst}) \times 10^{-6}$, which is consistent with the previous measurements and theoretical expectations. In addition, no significant $η\to e^+e^-$ signal is observed in the $e^+ e^-$ invariant mass spectrum, and an improved upper limit of ${\cal B}(η\to e^+ e^-) < 2.2 \times 10^{-7}$ is set at 90\% confidence level. △ Less

Submitted 7 December, 2025; originally announced December 2025.

Comments: 10 pages, 6 figures

Multi-band ALMA Polarization Observations of BHB07-11 Reveal Aligned Dust Grains in Complex Spiral Arm Structures

Authors: Austen Fourkas, Leslie W. Looney, Zhe-Yu Daniel Lin, Martin Radecki, Zhi-Yun Li, John J. Tobin, Ian W. Stephens, Manuel Fernández-López, Haifeng Yang, Woojin Kwon, Rachel Harrison

Abstract: Polarization-mode observations from the Atacama Large Millimeter/submillimeter Array (ALMA) are powerful tools for studying the dust grain populations in circumstellar disks. Many sources exhibit polarization signatures consistent with aligned dust grains, yet the physical origin of this alignment remains uncertain. One such source is BHB07-11, a Class I protobinary object in the Pipe Nebula with… ▽ More Polarization-mode observations from the Atacama Large Millimeter/submillimeter Array (ALMA) are powerful tools for studying the dust grain populations in circumstellar disks. Many sources exhibit polarization signatures consistent with aligned dust grains, yet the physical origin of this alignment remains uncertain. One such source is BHB07-11, a Class I protobinary object in the Pipe Nebula with complex spiral arm structures in its circumbinary disk. While magnetic fields are often invoked to explain grain alignment in the interstellar medium, the contrasting conditions in circumstellar disk environments demand further investigation into grain alignment mechanisms. To determine BHB07-11's dominant polarization mechanism, we leverage ALMA polarization-mode dust continuum observations in Bands 3 ($λ$=3.1 mm), 6 ($λ$=1.3 mm), and 7 ($λ$=0.87 mm), in combination with high-resolution dust continuum and spectral line observations in Band 6. Observed polarization vectors in each band are consistent with emission from aligned grains and follow the structure of the spiral arms as shown in the high-resolution observations. Given the relationship between the observed polarization vector orientation and the spiral arms, we find that the polarization morphology is most consistent with grains aligned through a relative velocity flow between gas and dust in the spiral arms, as envisioned in the recently developed badminton birdie-like alignment mechanism, rather than alignment with a magnetic field or other known alignment mechanisms. △ Less

Submitted 5 December, 2025; originally announced December 2025.

Comments: Accepted to ApJ, 30 pages, 21 figures

Evidence for the semileptonic decays $Λ_c^{+} \to Σ^{\pm} π^{\mp} e^+ ν_e$

Authors: BESIII Collaboration, M. Ablikim, M. N. Achasov, P. Adlarson, X. C. Ai, R. Aliberti, A. Amoroso, Q. An, Y. Bai, O. Bakina, Y. Ban, H. -R. Bao, V. Batozskaya, K. Begzsuren, N. Berger, M. Berlowski, M. Bertani, D. Bettoni, F. Bianchi, E. Bianco, A. Bortone, I. Boyko, R. A. Briere, A. Brueggemann, H. Cai , et al. (679 additional authors not shown)

Abstract: Using $4.5\, fb^{-1}$ of $e^+e^-$ collision data collected by the BESIII detector at center-of-mass energies between $4.600$ and $4.699\,GeV$, we search for the semileptonic decays $Λ_c^{+} \to Σ^{+} π^{-} e^+ ν_e$ and $Λ_c^{+} \to Σ^{-} π^{+} e^+ ν_e$ for the first time. Assuming their branching fractions are equal under isospin symmetry, evidence for $Λ_c^{+} \to Σ^{\pm} π^{\mp} e^+ ν_e$ is repo… ▽ More Using $4.5\, fb^{-1}$ of $e^+e^-$ collision data collected by the BESIII detector at center-of-mass energies between $4.600$ and $4.699\,GeV$, we search for the semileptonic decays $Λ_c^{+} \to Σ^{+} π^{-} e^+ ν_e$ and $Λ_c^{+} \to Σ^{-} π^{+} e^+ ν_e$ for the first time. Assuming their branching fractions are equal under isospin symmetry, evidence for $Λ_c^{+} \to Σ^{\pm} π^{\mp} e^+ ν_e$ is reported with a significance of $3.6σ$. The corresponding branching fraction is measured to be $\mathcal{B}(Λ_c^{+} \to Σ^\pmπ^\mp e^+ν_e) = (7.7^{+2.5}_{-2.3_{\rm stat.}}\pm1.3_{\rm syst.})\times 10^{-4}$, which is consistent with quark model predictions within two standard deviations. △ Less

Submitted 4 December, 2025; originally announced December 2025.

ARM-Thinker: Reinforcing Multimodal Generative Reward Models with Agentic Tool Use and Visual Reasoning

Authors: Shengyuan Ding, Xinyu Fang, Ziyu Liu, Yuhang Zang, Yuhang Cao, Xiangyu Zhao, Haodong Duan, Xiaoyi Dong, Jianze Liang, Bin Wang, Conghui He, Dahua Lin, Jiaqi Wang

Abstract: Reward models are critical for aligning vision-language systems with human preferences, yet current approaches suffer from hallucination, weak visual grounding, and an inability to use tools for verification, limiting their reliability on complex multimodal reasoning tasks. We present ARM-Thinker, an A}gentic multimodal Reward Model that autonomously invokes external tools (e.g., image cropping, d… ▽ More Reward models are critical for aligning vision-language systems with human preferences, yet current approaches suffer from hallucination, weak visual grounding, and an inability to use tools for verification, limiting their reliability on complex multimodal reasoning tasks. We present ARM-Thinker, an A}gentic multimodal Reward Model that autonomously invokes external tools (e.g., image cropping, doc page retrieval) to ground judgments in verifiable evidence, replacing static, non-interactive reward scoring. This enables the model to verify fine-grained visual details, cross-reference multi-page evidence, and validate reasoning claims, which are capabilities absent in existing reward models. We train ARM-Thinker with multi-stage reinforcement learning, jointly optimizing tool-calling decisions and judgment accuracy. To evaluate agentic reward modeling, we introduce ARMBench-VL, comprising three benchmarks that assess fine-grained visual grounding (image-level tools), multi-page document understanding (retrieval tools), and instruction following (text-level verification). ARM-Thinker achieves +16.2% average improvement on reward modeling benchmarks, +9.6% on tool-use tasks, and outperforms baselines on multimodal math and logical reasoning benchmarks. Our results demonstrate that agentic capabilities significantly enhance both accuracy and interpretability of reward models. △ Less

Submitted 4 December, 2025; originally announced December 2025.

Study of the reaction $Ξ^{0}n\rightarrowΛΛX$ using $Ξ^{0}$-nucleus scattering

Authors: BESIII Collaboration, M. Ablikim, M. N. Achasov, P. Adlarson, X. C. Ai, R. Aliberti, A. Amoroso, Q. An, Y. Bai, O. Bakina, Y. Ban, H. -R. Bao, V. Batozskaya, K. Begzsuren, N. Berger, M. Berlowski, M. Bertani, D. Bettoni, F. Bianchi, E. Bianco, A. Bortone, I. Boyko, R. A. Briere, A. Brueggemann, H. Cai , et al. (707 additional authors not shown)

Abstract: Using $(10087\pm44)\times10^{6}$$J/ψ$ events collected with the BESIII detector operating at the BEPCII storage ring in $2009$, $2012$, $2018$, and $2019$, we perform a search for the reaction $Ξ^0n\rightarrowΛΛX$, where $X$ denotes any additional final particles. Given the highly suppressed phase space for producing extra pions, the $X$ consists of either nothing or a photon, corresponding to the… ▽ More Using $(10087\pm44)\times10^{6}$$J/ψ$ events collected with the BESIII detector operating at the BEPCII storage ring in $2009$, $2012$, $2018$, and $2019$, we perform a search for the reaction $Ξ^0n\rightarrowΛΛX$, where $X$ denotes any additional final particles. Given the highly suppressed phase space for producing extra pions, the $X$ consists of either nothing or a photon, corresponding to the processes $Ξ^0 n \rightarrow ΛΛ$ and $Ξ^{0}n\rightarrowΛΣ^0\rightarrowΛΛγ$. The $Ξ^0$ comes from the decay of $J/ψ\rightarrowΞ^0\barΞ^0$, while the neutron originates from material of the beam pipe. A signal is observed for the first time with a statistical significance of 6.4$σ$. The cross section for the reaction $Ξ^0+{^9\rm{Be}}\rightarrowΛ+Λ+X$ is measured to be $(43.6\pm10.5_{\text{stat}}\pm11.1_{\text{syst}})$ mb at $P_{Ξ^0}\approx0.818$ GeV/$c$, where the first uncertainty is statistical and the second systematic. No significant $H$-dibaryon signal is observed in the $ΛΛ$ final state. △ Less

Submitted 4 December, 2025; originally announced December 2025.

Comments: 9 pages, 3 figures

Design and Performance Simulation of the Electromagnetic Calorimeter at EicC

Authors: Ye Tian, Souvik Maity, Jingyu Li, Yuancai Wu, Shan Sha, Yutie Liang, Aiqiang Guo, Yuxiang Zhao, Dexu Lin

Abstract: The electromagnetic calorimeter (ECAL) is a key detector component for precise electron and photon measurements in electron-ion collision experiments. At the Electron-Ion Collider in China (EicC), high-performance calorimetry is essential for exploring the internal structure of nucleons and studying the dynamics of quarks and gluons within quantum chromodynamics (QCD). This paper presents the opti… ▽ More The electromagnetic calorimeter (ECAL) is a key detector component for precise electron and photon measurements in electron-ion collision experiments. At the Electron-Ion Collider in China (EicC), high-performance calorimetry is essential for exploring the internal structure of nucleons and studying the dynamics of quarks and gluons within quantum chromodynamics (QCD). This paper presents the optimized design and performance simulation of the EicC ECAL system. The ECAL consists of three specialized sections tailored to distinct detection environments: (1) an electron-Endcap employing high-resolution pure Cesium Iodide (pCsI) crystals, (2) a central barrel, and (3) an ion-Endcap, both adopting a cost-effective Shashlik-style sampling calorimeter with improved light yield. Each segment's geometry and material composition have been systematically optimized through Geant4 simulations to achieve excellent energy and position resolutions as well as strong electron-pion discrimination. The simulated performance indicates that the ECAL can achieve energy resolutions of 2 percent divided by sqrt(E) for pCsI crystals and 5 percent divided by sqrt(E) for Shashlik modules, meeting the design goals of the EicC detector. △ Less

Submitted 10 December, 2025; v1 submitted 3 December, 2025; originally announced December 2025.

Measurement of the hyperon weak radiative decay $Ξ^0\toγΣ^0$ at BESIII

Authors: BESIII Collaboration, M. Ablikim, M. N. Achasov, P. Adlarson, X. C. Ai, R. Aliberti, A. Amoroso, M. R. An, Q. An, Y. Bai, O. Bakina, I. Balossino, Y. Ban, V. Batozskaya, K. Begzsuren, N. Berger, M. Berlowski, M. Bertani, D. Bettoni, F. Bianchi, E. Bianco, A. Bortone, I. Boyko, R. A. Briere, A. Brueggemann , et al. (604 additional authors not shown)

Abstract: The hyperon weak radiative decay $Ξ^0\toγΣ^0$ is measured via the process $J/ψ\toΞ^0\barΞ^0$. The absolute branching fraction of $Ξ^0\toγΣ^0$ is determined to be $(3.69\pm 0.21_{\text{stat}}\pm0.12_{\text{syst}})\times 10^{-3}$, based on $(10.087\pm 0.044)\times 10^{9}$ $J/ψ$ events collected with the BESIII detector operating at the BEPCII collider. The decay asymmetry parameter is measured, with… ▽ More The hyperon weak radiative decay $Ξ^0\toγΣ^0$ is measured via the process $J/ψ\toΞ^0\barΞ^0$. The absolute branching fraction of $Ξ^0\toγΣ^0$ is determined to be $(3.69\pm 0.21_{\text{stat}}\pm0.12_{\text{syst}})\times 10^{-3}$, based on $(10.087\pm 0.044)\times 10^{9}$ $J/ψ$ events collected with the BESIII detector operating at the BEPCII collider. The decay asymmetry parameter is measured, with a complete angular analysis of the cascade decay chain, to be $α_γ = -0.807\pm 0.095_{\text{stat}}\pm 0.011_{\text{syst}}$. △ Less

Submitted 3 December, 2025; originally announced December 2025.

Ensemble Privacy Defense for Knowledge-Intensive LLMs against Membership Inference Attacks

Authors: Haowei Fu, Bo Ni, Han Xu, Kunpeng Liu, Dan Lin, Tyler Derr

Abstract: Retrieval-Augmented Generation (RAG) and Supervised Finetuning (SFT) have become the predominant paradigms for equipping Large Language Models (LLMs) with external knowledge for diverse, knowledge-intensive tasks. However, while such knowledge injection improves performance, it also exposes new attack surfaces. Membership Inference Attacks (MIAs), which aim to determine whether a given data sample… ▽ More Retrieval-Augmented Generation (RAG) and Supervised Finetuning (SFT) have become the predominant paradigms for equipping Large Language Models (LLMs) with external knowledge for diverse, knowledge-intensive tasks. However, while such knowledge injection improves performance, it also exposes new attack surfaces. Membership Inference Attacks (MIAs), which aim to determine whether a given data sample was included in a model's training set, pose serious threats to privacy and trust in sensitive domains. To this end, we first systematically evaluate the vulnerability of RAG- and SFT-based LLMs to various MIAs. Then, to address the privacy risk, we further introduce a novel, model-agnostic defense framework, Ensemble Privacy Defense (EPD), which aggregates and evaluates the outputs of a knowledge-injected LLM, a base LLM, and a dedicated judge model to enhance resistance against MIAs. Comprehensive experiments show that, on average, EPD reduces MIA success by up to 27.8\% for SFT and 526.3\% for RAG compared to inference-time baseline, while maintaining answer quality. △ Less

Submitted 1 December, 2025; originally announced December 2025.

ViSAudio: End-to-End Video-Driven Binaural Spatial Audio Generation

Authors: Mengchen Zhang, Qi Chen, Tong Wu, Zihan Liu, Dahua Lin

Abstract: Despite progress in video-to-audio generation, the field focuses predominantly on mono output, lacking spatial immersion. Existing binaural approaches remain constrained by a two-stage pipeline that first generates mono audio and then performs spatialization, often resulting in error accumulation and spatio-temporal inconsistencies. To address this limitation, we introduce the task of end-to-end b… ▽ More Despite progress in video-to-audio generation, the field focuses predominantly on mono output, lacking spatial immersion. Existing binaural approaches remain constrained by a two-stage pipeline that first generates mono audio and then performs spatialization, often resulting in error accumulation and spatio-temporal inconsistencies. To address this limitation, we introduce the task of end-to-end binaural spatial audio generation directly from silent video. To support this task, we present the BiAudio dataset, comprising approximately 97K video-binaural audio pairs spanning diverse real-world scenes and camera rotation trajectories, constructed through a semi-automated pipeline. Furthermore, we propose ViSAudio, an end-to-end framework that employs conditional flow matching with a dual-branch audio generation architecture, where two dedicated branches model the audio latent flows. Integrated with a conditional spacetime module, it balances consistency between channels while preserving distinctive spatial characteristics, ensuring precise spatio-temporal alignment between audio and the input video. Comprehensive experiments demonstrate that ViSAudio outperforms existing state-of-the-art methods across both objective metrics and subjective evaluations, generating high-quality binaural audio with spatial immersion that adapts effectively to viewpoint changes, sound-source motion, and diverse acoustic environments. Project website: https://kszpxxzmc.github.io/ViSAudio-project. △ Less

Submitted 2 December, 2025; originally announced December 2025.

Tidal disruption and evaporation of rubble-pile and monolithic bodies as a source of flaring activity in Sgr A^\star$

Authors: Wen-Han Zhou, Yun Zhang, Jiamu Huang, Douglas N. C. Lin

Abstract: Sgr A*, the supermassive black hole at the center of the Milky Way, exhibits frequent short-duration flares with luminosity greater than 1e34 erg/s across multiple wavelengths. The origin of the flares is still unknown. We revisited the role of small planetary bodies, originally from the stellar disk, and their tidally disrupted fragments as a source of flaring activity in Sgr A*. We refined previ… ▽ More Sgr A*, the supermassive black hole at the center of the Milky Way, exhibits frequent short-duration flares with luminosity greater than 1e34 erg/s across multiple wavelengths. The origin of the flares is still unknown. We revisited the role of small planetary bodies, originally from the stellar disk, and their tidally disrupted fragments as a source of flaring activity in Sgr A*. We refined previous models by incorporating material strength constraints on the tidal disruption limit and by evaluating the evaporation dynamics of the resulting fragments. We analyzed the tidal fragmentation and gas-induced fragmentation of small planetary bodies with rubble-pile and monolithic structures. Using constraints from recent space missions (e.g., NASA OSIRIS-REx and JAXA Hayabusa2), we estimated the survivability of fragments under aerodynamic heating and computed their expected luminosity from ablation, modeled as fireball flares analogous to meteor events. We find that planetary fragments can approach as close as 8 gravitational radii, consistent with observed flare locations. The fireball model yields luminosities from 1e34 to 1e36 erg/s for fragments whose parent bodies are a few kilometers in size. The derived flare frequency vs. luminosity distribution follows a power law with index 1.83, in agreement with observed values (1.65 - 1.9), while the flare duration scales as L^(-1/3), consistent with observations. We consider the young stars around Sgr A* as the planetary reservoir. Given a small-body population analogous in mass to the primordial Kuiper belt and the common existence of close-in super-Earths and long-period Neptunes, we show that this planetary reservoir can supply the observed flares. △ Less

Submitted 30 November, 2025; originally announced December 2025.

Comments: Accepted to A&A. 9 pages, 3 figures

Outward Migration of a Gas Accreting Planet: A Semi-Analytical Formula

Authors: Shigeru Ida, Ya-Ping Li, Jun-Peng Pan, Yi-Xian Chen, Douglas N. C. Lin

Abstract: Type II orbital migration is a key process to regulate the mass and semimajor axis distribution of exoplanetary giant planets. The conventional formula of type II migration generally predicts too rapid inward migration to reconcile with the observed pile-up of gas giant beyond 1 au. Analyzing the recent high-resolution hydrodynamical simulations by Li et al. (2024) and Pan et al. (2025) that show… ▽ More Type II orbital migration is a key process to regulate the mass and semimajor axis distribution of exoplanetary giant planets. The conventional formula of type II migration generally predicts too rapid inward migration to reconcile with the observed pile-up of gas giant beyond 1 au. Analyzing the recent high-resolution hydrodynamical simulations by Li et al. (2024) and Pan et al. (2025) that show robust outward migration of a gas accreting planet, we here clarify the condition for the outward migration to occur and derive a general semi-analytical formula that can be applied for broad range of planet mass and disk conditions. The striking outward migration is caused by azimuthal asymmetry in corotation torque exerted from cicumplanetary disk regions (connecting to horseshoe flow) that is produced by the planetary gas accretion, while the conventional inward migration model is based on radial asymmetry in the torques from the circumstellar protoplanetry disk. We found that the azimuthal asymmetry dominates and the migration is outward, when the gap depth defined by the surface density reduction factor of $1/(1+K')$ is in the range of $0.03 \lesssim K' \lesssim 50$. Using simple models with the new formula, we demonstrate that the outward migration plays an important role in shaping the mass and semimajor axis distribution of gas giants. The concurrent dependence of planets' accretion rate and migration direction on their masses and disk properties potentially reproduces the observed pile-up of exoplanetary gas giants beyond 1 au, although more detailed planet population synthesis calculations are needed in the future. △ Less

Submitted 28 November, 2025; originally announced December 2025.

Comments: Accepted for publication in The Astrophysical Journal, 12 pages, 5 figures

Study of the reactions $\bar{n} p \to 2π^{+}π^{-}$, $2π^{+}π^{-}π^{0}$, and $2π^{+}π^{-}2π^{0}$ using $J/ψ\to p π^{-}\bar{n}$

Authors: BESIII Collaboration, M. Ablikim, M. N. Achasov, P. Adlarson, X. C. Ai, R. Aliberti, A. Amoroso, Q. An, Y. Bai, O. Bakina, Y. Ban, H. -R. Bao, X. L. Bao, V. Batozskaya, K. Begzsuren, N. Berger, M. Berlowski, M. B. Bertani, D. Bettoni, F. Bianchi, E. Bianco, A. Bortone, I. Boyko, R. A. Briere, A. Brueggemann , et al. (687 additional authors not shown)

Abstract: We report an experimental investigation of the reactions $\bar{n} p \to 2π^{+}π^{-}$, $\bar{n} p \to 2π^{+}π^{-}π^{0}$, and $\bar{n} p \to 2π^{+}π^{-}2π^{0}$ using $(10.087 \pm 0.044) \times 10^{9}$ $J/ψ$ events collected with the BESIII detector at the BEPCII storage ring. The antineutron ($\bar{n}$) is produced in the decay $J/ψ\to p π^{-} \bar{n}$ with studied momentum from 200~MeV/$c$ to 1174~… ▽ More We report an experimental investigation of the reactions $\bar{n} p \to 2π^{+}π^{-}$, $\bar{n} p \to 2π^{+}π^{-}π^{0}$, and $\bar{n} p \to 2π^{+}π^{-}2π^{0}$ using $(10.087 \pm 0.044) \times 10^{9}$ $J/ψ$ events collected with the BESIII detector at the BEPCII storage ring. The antineutron ($\bar{n}$) is produced in the decay $J/ψ\to p π^{-} \bar{n}$ with studied momentum from 200~MeV/$c$ to 1174~MeV/$c$, while the target proton originates from the hydrogen nuclei in the cooling oil of the beam pipe. This novel method pioneers the study of $\bar{n}$-nucleon interactions at an $e^{+}e^{-}$ collider, providing the first experimental data for $\bar{n}$ momenta exceeding 800~MeV/$c$. △ Less

Submitted 26 November, 2025; originally announced November 2025.

Planet Migration in Protoplanetary Disks with Rims

Authors: Zhuoya Cao, Ya-Ping Li, Douglas N. C. Lin, Shude Mao

Abstract: Complex structures, including sharp edges, rings and gaps, have been commonly observed in protoplanetary disks with or without planetary candidates. Here we consider the possibility that they are the intrinsic consequences of angular momentum transfer mechanisms, and investigate how they may influence the dynamical evolution of embedded planets. With the aid of numerical hydrodynamic simulations,… ▽ More Complex structures, including sharp edges, rings and gaps, have been commonly observed in protoplanetary disks with or without planetary candidates. Here we consider the possibility that they are the intrinsic consequences of angular momentum transfer mechanisms, and investigate how they may influence the dynamical evolution of embedded planets. With the aid of numerical hydrodynamic simulations, we show that gas giants have a tendency to migrate away from sharp edges, whereas super-Earths embedded in the annuli tend to be retained. This implies that, observationally, Jupiters are preferentially detected in dark rings (gaps), whereas super-Earths tend to be found in bright rings (density bumps). Moreover, planets' tidal torque provide, not necessarily predominant, feedback on the surface density profile. This tendency implies that Jupiter's gap-opening process deepens and widens the density gap associated with the dark ring, while super-Earths can be halted by steep surface density gradient near the disk or ring boundaries. 13Hence, we expect there would be a desert for super-Earths in the surface density gap. △ Less

Submitted 26 November, 2025; originally announced November 2025.

Comments: 13 pages, 7 figures, Submitted to APJ

Concurrent Accretion and Migration of Giant Planets in their Natal Disks with Consistent Accretion Torque (II): Parameter Survey and Condition for Outward Migration

Authors: JunPeng Pan, Ya-Ping Li, Yi-Xian Chen, Shigeru Ida, Douglas N. C. Lin

Abstract: Migration typically occurs during the formation of planets and is closely linked to the planetary formation process. In classical theories of non-accreting planetary migration, both type I and type II migration typically result in inward migration, which is hard to align with the architecture of the planetary systems.In this work, we conduct systematic, high-resolution 3D/2D numerical hydrodynamic… ▽ More Migration typically occurs during the formation of planets and is closely linked to the planetary formation process. In classical theories of non-accreting planetary migration, both type I and type II migration typically result in inward migration, which is hard to align with the architecture of the planetary systems.In this work, we conduct systematic, high-resolution 3D/2D numerical hydrodynamic simulations to investigate the migration of an accreting planet. Under different disk conditions, we compared the dynamical evolution of planets with different planet-to-star mass ratios. We find that accretion of planets can significantly diminish the inward migration tendency of planets, or even change the migration direction. The migration of low-/high-mass planets is classified as Type I/II inward migration, respectively, while intermediate-mass planets, which have the strongest accretion, show an outward migration trend. We confirm that the outward migration is mainly attributed to the positive torque from the azimuthal asymmetric structures around the accreting planet, similar to Li et al. (2024). The termination of planetary mass growth is thus synonymous with the transition from outward to inward migration. For the high viscosity $α=0.04$ and disk aspect ratio height $h_0=0.05$ cases, the mass ratio range for planetary outward migration is $1\times10^{-4}\lesssim q\lesssim4\times10^{-3}$. For the low viscosity case with $α=0.001$, and/or the low disk aspect ratio cases $h_0=0.03$, the mass ratio range for the outward migration will shift toward the lower end. Our parameter survey reveals that a simple gap opening parameter determines the outward migration condition; details of the analytical interpretation are presented in Ida et al. (2025). △ Less

Submitted 26 November, 2025; originally announced November 2025.

Comments: 14 pages, 7 figures, 1 table. Accepted for publication in The Astrophysical Journal

ConsistCompose: Unified Multimodal Layout Control for Image Composition

Authors: Xuanke Shi, Boxuan Li, Xiaoyang Han, Zhongang Cai, Lei Yang, Dahua Lin, Quan Wang

Abstract: Unified multimodal models that couple visual understanding with image generation have advanced rapidly, yet most systems still focus on visual grounding-aligning language with image regions-while their generative counterpart, linguistic-embedded layout-grounded generation (LELG) for layout-controllable multi-instance generation, remains underexplored and limits precise compositional control. We pr… ▽ More Unified multimodal models that couple visual understanding with image generation have advanced rapidly, yet most systems still focus on visual grounding-aligning language with image regions-while their generative counterpart, linguistic-embedded layout-grounded generation (LELG) for layout-controllable multi-instance generation, remains underexplored and limits precise compositional control. We present ConsistCompose, a unified multimodal framework that embeds layout coordinates directly into language prompts, enabling layout-controlled multi-instance image generation from Interleaved Image-Text within a single generative interface. We further construct ConsistCompose3M, a 3.4M multi-instance generation dataset with layout and identity annotations (2.6M text-guided and 0.8M image-guided data pairs) that provides large-scale supervision for layout-conditioned generation. Within this framework, LELG is instantiated through instance-coordinate binding prompts and coordinate-aware classifier-free guidance, which translate linguistic layout cues into precise spatial control without task-specific branches. Experiments on COCO-Position and MS-Bench show that ConsistCompose substantially improves spatial accuracy over layout-controlled baselines while preserving identity fidelity and competitive general multimodal understanding, establishing a unified paradigm for layout-controllable multimodal image generation. △ Less

Submitted 11 December, 2025; v1 submitted 23 November, 2025; originally announced November 2025.

Comments: 22 pages, 17 figures

Which active galaxies might be neutrino emitters?

Authors: Shuying Zhou, Mouyuan Sun, Guobin Mou, Da-bin Lin, Tong Liu, Ming-Xuan Lu, Yongquan Xue

Abstract: The IceCube Neutrino Observatory has identified several individual neutrino emitters associated with supermassive black hole accretion phenomena, including blazars, tidal disruption events, and, unexpectedly, Seyfert galaxies. A key open question is which types of active galactic nuclei (AGNs) are most likely to be neutrino emitters. Here we show that high-confidence extragalactic neutrino emitter… ▽ More The IceCube Neutrino Observatory has identified several individual neutrino emitters associated with supermassive black hole accretion phenomena, including blazars, tidal disruption events, and, unexpectedly, Seyfert galaxies. A key open question is which types of active galactic nuclei (AGNs) are most likely to be neutrino emitters. Here we show that high-confidence extragalactic neutrino emitters tend not only to have higher hard X-ray fluxes but also to be more variable in mid-infrared (MIR) than other AGNs in the \textit{Swift} BAT AGN Spectroscopic Survey. MIR variations effectively trace long-term fluctuations in AGN accretion disks and/or jets. In addition to the role of X-ray flux emphasized in previous studies, we speculate that long-term central engine fluctuations may also be critical for neutrino production. This hypothesis may inform IceCube neutrino-electromagnetic counterpart association studies and provide new insights into cosmic ray acceleration sites. First, the observed neutrinos are unlikely to originate from AGN host galaxies or from interactions between large-scale (dozens of parsecs) winds/outflows and the surrounding interstellar medium. Second, if neutrinos are produced in the X-ray corona, the corona should exhibit strong magnetic turbulence dissipation or magnetic reconnection whose rate changes substantially on timescales of years. Third, the relativistic jets of blazar neutrino emitters may be intrinsically unstable over years. Finally, if neutrinos are related to interactions between small-scale winds/outflows and torus clouds, such winds/outflows must be highly episodic. △ Less

Submitted 20 November, 2025; originally announced November 2025.

Comments: 11 pages, 3 figures, 1 table. accepted to ApJ. Comments welcome

AICC: Parse HTML Finer, Make Models Better -- A 7.3T AI-Ready Corpus Built by a Model-Based HTML Parser

Authors: Ren Ma, Jiantao Qiu, Chao Xu, Pei Chu, Kaiwen Liu, Pengli Ren, Yuan Qu, Jiahui Peng, Linfeng Hou, Mengjie Liu, Lindong Lu, Wenchang Ning, Jia Yu, Rui Min, Jin Shi, Haojiong Chen, Peng Zhang, Wenjian Zhang, Qian Jiang, Zengjie Hu, Guoqiang Yang, Zhenxiang Li, Fukai Shang, Runyuan Ma, Chenlin Su , et al. (4 additional authors not shown)

Abstract: While web data quality is crucial for large language models, most curation efforts focus on filtering and deduplication,treating HTML-to-text extraction as a fixed pre-processing step. Existing web corpora rely on heuristic-based extractors like Trafilatura, which struggle to preserve document structure and frequently corrupt structured elements such as formulas, codes, and tables. We hypothesize… ▽ More While web data quality is crucial for large language models, most curation efforts focus on filtering and deduplication,treating HTML-to-text extraction as a fixed pre-processing step. Existing web corpora rely on heuristic-based extractors like Trafilatura, which struggle to preserve document structure and frequently corrupt structured elements such as formulas, codes, and tables. We hypothesize that improving extraction quality can be as impactful as aggressive filtering strategies for downstream performance. We introduce MinerU-HTML, a novel extraction pipeline that reformulates content extraction as a sequence labeling problem solved by a 0.6B-parameter language model. Unlike text-density heuristics, MinerU-HTML leverages semantic understanding and employs a two-stage formatting pipeline that explicitly categorizes semantic elements before converting to Markdown. Crucially, its model-based approach is inherently scalable, whereas heuristic methods offer limited improvement pathways. On MainWebBench, our benchmark of 7,887 annotated web pages, MinerU-HTML achieves 81.8\% ROUGE-N F1 compared to Trafilatura's 63.6\%, with exceptional structured element preservation (90.9\% for code blocks, 94.0\% for formulas). Using MinerU-HTML, we construct AICC (AI-ready Common Crawl), a 7.3-trillion token multilingual corpus from two Common Crawl snapshots. In controlled pretraining experiments where AICC and Trafilatura-extracted TfCC undergo identical filtering, models trained on AICC (62B tokens) achieve 50.8\% average accuracy across 13 benchmarks, outperforming TfCC by 1.08pp-providing direct evidence that extraction quality significantly impacts model capabilities. AICC also surpasses RefinedWeb and FineWeb on key benchmarks. We publicly release MainWebBench, MinerU-HTML, and AICC, demonstrating that HTML extraction is a critical, often underestimated component of web corpus construction. △ Less

Submitted 26 November, 2025; v1 submitted 20 November, 2025; originally announced November 2025.

Search for the charmonium weak decay $J/ψ\to\bar{D}^0\bar{K}^{*0}+{\rm c.c.}$

Authors: BESIII Collaboration, M. Ablikim, M. N. Achasov, P. Adlarson, X. C. Ai, R. Aliberti, A. Amoroso, Q. An, Y. Bai, O. Bakina, Y. Ban, H. -R. Bao, V. Batozskaya, K. Begzsuren, N. Berger, M. Berlowski, M. Bertani, D. Bettoni, F. Bianchi, E. Bianco, A. Bortone, I. Boyko, R. A. Briere, A. Brueggemann, H. Cai , et al. (706 additional authors not shown)

Abstract: Based on a sample of $(10087\pm44)\times10^6$ $J/ψ$ events collected at the center-of-mass energy $\sqrt{s}$ = 3.0969 GeV with the BESIII detector, we search for the charmonium rare weak decay $J/ψ\to\bar{D}^0\bar{K}^{*0}+{\rm c.c.}$. No significant signal is observed, and the upper limit on its decay branching fraction at the 90% confidence level is set as $1.9\times10^{-7}$, improving the sensit… ▽ More Based on a sample of $(10087\pm44)\times10^6$ $J/ψ$ events collected at the center-of-mass energy $\sqrt{s}$ = 3.0969 GeV with the BESIII detector, we search for the charmonium rare weak decay $J/ψ\to\bar{D}^0\bar{K}^{*0}+{\rm c.c.}$. No significant signal is observed, and the upper limit on its decay branching fraction at the 90% confidence level is set as $1.9\times10^{-7}$, improving the sensitivity of the previous best limit by an order of magnitude. △ Less

Submitted 20 November, 2025; originally announced November 2025.

Think Visually, Reason Textually: Vision-Language Synergy in ARC

Authors: Beichen Zhang, Yuhang Zang, Xiaoyi Dong, Yuhang Cao, Haodong Duan, Dahua Lin, Jiaqi Wang

Abstract: Abstract reasoning from minimal examples remains a core unsolved problem for frontier foundation models such as GPT-5 and Grok 4. These models still fail to infer structured transformation rules from a handful of examples, which is a key hallmark of human intelligence. The Abstraction and Reasoning Corpus for Artificial General Intelligence (ARC-AGI) provides a rigorous testbed for this capability… ▽ More Abstract reasoning from minimal examples remains a core unsolved problem for frontier foundation models such as GPT-5 and Grok 4. These models still fail to infer structured transformation rules from a handful of examples, which is a key hallmark of human intelligence. The Abstraction and Reasoning Corpus for Artificial General Intelligence (ARC-AGI) provides a rigorous testbed for this capability, demanding conceptual rule induction and transfer to novel tasks. Most existing methods treat ARC-AGI as a purely textual reasoning task, overlooking the fact that humans rely heavily on visual abstraction when solving such puzzles. However, our pilot experiments reveal a paradox: naively rendering ARC-AGI grids as images degrades performance due to imprecise rule execution. This leads to our central hypothesis that vision and language possess complementary strengths across distinct reasoning stages: vision supports global pattern abstraction and verification, whereas language specializes in symbolic rule formulation and precise execution. Building on this insight, we introduce two synergistic strategies: (1) Vision-Language Synergy Reasoning (VLSR), which decomposes ARC-AGI into modality-aligned subtasks; and (2) Modality-Switch Self-Correction (MSSC), which leverages vision to verify text-based reasoning for intrinsic error correction. Extensive experiments demonstrate that our approach yields up to a 4.33\% improvement over text-only baselines across diverse flagship models and multiple ARC-AGI tasks. Our findings suggest that unifying visual abstraction with linguistic reasoning is a crucial step toward achieving generalizable, human-like intelligence in future foundation models. Source code is released at https://github.com/InternLM/ARC-VL. △ Less

Submitted 26 November, 2025; v1 submitted 19 November, 2025; originally announced November 2025.

Search for the lepton number violating process $Ξ^- \rightarrow Σ^+ e^- e^- +c.c.$

Authors: BESIII Collaboration, M. Ablikim, M. N. Achasov, P. Adlarson, X. C. Ai, R. Aliberti, A. Amoroso, Q. An, Y. Bai, O. Bakina, Y. Ban, H. -R. Bao, X. L. Bao, V. Batozskaya, K. Begzsuren, N. Berger, M. Berlowski, M. B. Bertani, D. Bettoni, F. Bianchi, E. Bianco, A. Bortone, I. Boyko, R. A. Briere, A. Brueggemann , et al. (691 additional authors not shown)

Abstract: We present a search for the lepton number violating decay $Ξ^-\rightarrowΣ^+e^-e^- +c.c.$ with $(10087\pm44)\times10^6$ $J/ψ$ events collected by the BESIII detector at the BEPCII collider. Employing a blind analysis strategy, no significant signal is observed above the expected background yield. The upper limit on the branching fraction is determined to be… ▽ More We present a search for the lepton number violating decay $Ξ^-\rightarrowΣ^+e^-e^- +c.c.$ with $(10087\pm44)\times10^6$ $J/ψ$ events collected by the BESIII detector at the BEPCII collider. Employing a blind analysis strategy, no significant signal is observed above the expected background yield. The upper limit on the branching fraction is determined to be ${\rm Br}(Ξ^-\rightarrowΣ^+e^-e^-+c.c.)< 2.0\times10^{-5}$ at the $90\%$ confidence level. △ Less

Submitted 19 November, 2025; originally announced November 2025.

Gallant: Voxel Grid-based Humanoid Locomotion and Local-navigation across 3D Constrained Terrains

Authors: Qingwei Ben, Botian Xu, Kailin Li, Feiyu Jia, Wentao Zhang, Jingping Wang, Jingbo Wang, Dahua Lin, Jiangmiao Pang

Abstract: Robust humanoid locomotion requires accurate and globally consistent perception of the surrounding 3D environment. However, existing perception modules, mainly based on depth images or elevation maps, offer only partial and locally flattened views of the environment, failing to capture the full 3D structure. This paper presents Gallant, a voxel-grid-based framework for humanoid locomotion and loca… ▽ More Robust humanoid locomotion requires accurate and globally consistent perception of the surrounding 3D environment. However, existing perception modules, mainly based on depth images or elevation maps, offer only partial and locally flattened views of the environment, failing to capture the full 3D structure. This paper presents Gallant, a voxel-grid-based framework for humanoid locomotion and local navigation in 3D constrained terrains. It leverages voxelized LiDAR data as a lightweight and structured perceptual representation, and employs a z-grouped 2D CNN to map this representation to the control policy, enabling fully end-to-end optimization. A high-fidelity LiDAR simulation that dynamically generates realistic observations is developed to support scalable, LiDAR-based training and ensure sim-to-real consistency. Experimental results show that Gallant's broader perceptual coverage facilitates the use of a single policy that goes beyond the limitations of previous methods confined to ground-level obstacles, extending to lateral clutter, overhead constraints, multi-level structures, and narrow passages. Gallant also firstly achieves near 100% success rates in challenging scenarios such as stair climbing and stepping onto elevated platforms through improved end-to-end optimization. △ Less

Submitted 18 November, 2025; originally announced November 2025.

Scaling Spatial Intelligence with Multimodal Foundation Models

Authors: Zhongang Cai, Ruisi Wang, Chenyang Gu, Fanyi Pu, Junxiang Xu, Yubo Wang, Wanqi Yin, Zhitao Yang, Chen Wei, Qingping Sun, Tongxi Zhou, Jiaqi Li, Hui En Pang, Oscar Qian, Yukun Wei, Zhiqian Lin, Xuanke Shi, Kewang Deng, Xiaoyang Han, Zukai Chen, Xiangyu Fan, Hanming Deng, Lewei Lu, Liang Pan, Bo Li , et al. (4 additional authors not shown)

Abstract: Despite remarkable progress, multimodal foundation models still exhibit surprising deficiencies in spatial intelligence. In this work, we explore scaling up multimodal foundation models to cultivate spatial intelligence within the SenseNova-SI family, built upon established multimodal foundations including visual understanding models (i.e., Qwen3-VL and InternVL3) and unified understanding and gen… ▽ More Despite remarkable progress, multimodal foundation models still exhibit surprising deficiencies in spatial intelligence. In this work, we explore scaling up multimodal foundation models to cultivate spatial intelligence within the SenseNova-SI family, built upon established multimodal foundations including visual understanding models (i.e., Qwen3-VL and InternVL3) and unified understanding and generation models (i.e., Bagel). We take a principled approach to constructing high-performing and robust spatial intelligence by systematically curating SenseNova-SI-8M: eight million diverse data samples under a rigorous taxonomy of spatial capabilities. SenseNova-SI demonstrates unprecedented performance across a broad range of spatial intelligence benchmarks: 68.7% on VSI-Bench, 43.3% on MMSI, 85.6% on MindCube, 54.6% on ViewSpatial, and 50.1% on SITE, while maintaining strong general multimodal understanding (e.g., 84.9% on MMBench-En). More importantly, we analyze the impact of data scaling, discuss early signs of emergent generalization capabilities enabled by diverse data training, analyze the risk of overfitting and language shortcuts, present a preliminary study on spatial chain-of-thought reasoning, and validate the potential downstream application. SenseNova-SI is an ongoing project, and this report will be updated continuously. All newly trained multimodal foundation models are publicly released to facilitate further research in this direction. △ Less

Submitted 27 November, 2025; v1 submitted 17 November, 2025; originally announced November 2025.

Comments: Codebase: https://github.com/OpenSenseNova/SenseNova-SI; Models: https://huggingface.co/collections/sensenova/sensenova-si

CorrectAD: A Self-Correcting Agentic System to Improve End-to-end Planning in Autonomous Driving

Authors: Enhui Ma, Lijun Zhou, Tao Tang, Jiahuan Zhang, Junpeng Jiang, Zhan Zhang, Dong Han, Kun Zhan, Xueyang Zhang, XianPeng Lang, Haiyang Sun, Xia Zhou, Di Lin, Kaicheng Yu

Abstract: End-to-end planning methods are the de facto standard of the current autonomous driving system, while the robustness of the data-driven approaches suffers due to the notorious long-tail problem (i.e., rare but safety-critical failure cases). In this work, we explore whether recent diffusion-based video generation methods (a.k.a. world models), paired with structured 3D layouts, can enable a fully… ▽ More End-to-end planning methods are the de facto standard of the current autonomous driving system, while the robustness of the data-driven approaches suffers due to the notorious long-tail problem (i.e., rare but safety-critical failure cases). In this work, we explore whether recent diffusion-based video generation methods (a.k.a. world models), paired with structured 3D layouts, can enable a fully automated pipeline to self-correct such failure cases. We first introduce an agent to simulate the role of product manager, dubbed PM-Agent, which formulates data requirements to collect data similar to the failure cases. Then, we use a generative model that can simulate both data collection and annotation. However, existing generative models struggle to generate high-fidelity data conditioned on 3D layouts. To address this, we propose DriveSora, which can generate spatiotemporally consistent videos aligned with the 3D annotations requested by PM-Agent. We integrate these components into our self-correcting agentic system, CorrectAD. Importantly, our pipeline is an end-to-end model-agnostic and can be applied to improve any end-to-end planner. Evaluated on both nuScenes and a more challenging in-house dataset across multiple end-to-end planners, CorrectAD corrects 62.5% and 49.8% of failure cases, reducing collision rates by 39% and 27%, respectively. △ Less

Submitted 17 November, 2025; originally announced November 2025.

Lead-Free Europium Halide Perovskite Nanoplatelets

Authors: Sebastian Fernández, Divine Mbachu, Manchen Hu, Han Cui, William Michaels, Pournima Narayanan, Tyler K. Colenbrander, Qi Zhou, Da Lin, Ona Segura Lecina, Guosong Hong, Daniel N. Congreve

Abstract: Metal halide perovskites possess desirable optical, material, and electrical properties which have had substantial impact on next-generation optoelectronics. However, given the toxicity of lead, alternative lead-free perovskite semiconductors are needed. By fully replacing lead with rare-earth elements, one can simultaneously address toxicity concerns and achieve comparable optoelectronic performa… ▽ More Metal halide perovskites possess desirable optical, material, and electrical properties which have had substantial impact on next-generation optoelectronics. However, given the toxicity of lead, alternative lead-free perovskite semiconductors are needed. By fully replacing lead with rare-earth elements, one can simultaneously address toxicity concerns and achieve comparable optoelectronic performance. Here, we demonstrate the synthesis of two-dimensional europium halide perovskite nanoplatelets governed by the formula $\mathrm{L_{2}EuX_{4}}$ where L is an organic ligand and X is a halide anion. The structure, morphology, and composition of the nanoplatelets are confirmed by XRD, AFM, and XPS. Deep blue-emitting $\mathrm{PEA_{2}EuBr_{4}}$ perovskite nanoplatelets are synthesized in both the solution- and solid-states with photoluminescence emission centered at 446 nm and CIE color coordinates of (0.1515, 0.0327) and (0.1515, 0.0342), respectively. Additionally, near-ultraviolet $\mathrm{PEA_{2}EuCl_{4}}$ perovskite nanoplatelets are synthesized in both the solution- and solid-states with photoluminescence emission centered at 400 nm and 401 nm, respectively. Overall, europium halide perovskite nanoplatelets offer a lead-free alternative for deep blue, violet, and near-ultraviolet light emission $\unicode{x2013}$ charting new pathways for optoelectronics in this energy regime. △ Less

Submitted 13 November, 2025; originally announced November 2025.

Robust Resource Allocation via Competitive Subsidies

Authors: David X. Lin, Giannis Fikioris, Siddhartha Banerjee, Éva Tardos

Abstract: A canonical setting for non-monetary online resource allocation is one where agents compete over multiple rounds for a single item per round, with i.i.d. valuations and additive utilities across rounds. With $n$ symmetric agents, a natural benchmark for each agent is the utility realized by her favorite $1/n$-fraction of rounds; a line of work has demonstrated one can robustly guarantee each agent… ▽ More A canonical setting for non-monetary online resource allocation is one where agents compete over multiple rounds for a single item per round, with i.i.d. valuations and additive utilities across rounds. With $n$ symmetric agents, a natural benchmark for each agent is the utility realized by her favorite $1/n$-fraction of rounds; a line of work has demonstrated one can robustly guarantee each agent a constant fraction of this ideal utility, irrespective of how other agents behave. In particular, several mechanisms have been shown to be $1/2$-robust, and recent work established that repeated first-price auctions based on artificial credits have a robustness factor of $0.59$, which cannot be improved beyond $0.6$ using first-price and simple strategies. In contrast, even without strategic considerations, the best achievable factor is $1-1/e\approx 0.63$. In this work, we break the $0.6$ first-price barrier to get a new $0.625$-robust mechanism, which almost closes the gap to the non-strategic robustness bound. Surprisingly, we do so via a simple auction, where in each round, bidders decide if they ask for the item, and we allocate uniformly at random among those who ask. The main new ingredient is the idea of competitive subsidies, wherein we charge the winning agent an amount in artificial credits that decreases when fewer agents are bidding (specifically, when $k$ agents bid, then the winner pays proportional to $k/(k+1)$, varying the payment by a factor of 2 depending on the competition). Moreover, we show how it can be modified to get an equilibrium strategy with a slightly weaker robust guarantee of $5/(3e) \approx 0.61$ (and the optimal $1-1/e$ factor at equilibrium). Finally, we show that our mechanism gives the best possible bound under a wide class of auction-based mechanisms. △ Less

Submitted 27 November, 2025; v1 submitted 12 November, 2025; originally announced November 2025.

glitterin: Towards Replacing the Role of Lorenz-Mie Theory in Astronomy Using Neural Networks Trained on Light Scattering of Irregularly Shaped Grains

Authors: Zhe-Yu Daniel Lin, Alycia J. Weinberger, Evgenij Zubko, Jessica A. Arnold, Gorden Videen

Abstract: Light scattering by dust particles is often modeled assuming the dust is spherical for numerical simplicity and speed. However, real dust particles have highly irregular morphologies that significantly affect their scattering properties. We have developed glitterin, a neural network trained to predict light scattering from irregularly shaped dust grains, offering a computationally efficient altern… ▽ More Light scattering by dust particles is often modeled assuming the dust is spherical for numerical simplicity and speed. However, real dust particles have highly irregular morphologies that significantly affect their scattering properties. We have developed glitterin, a neural network trained to predict light scattering from irregularly shaped dust grains, offering a computationally efficient alternative to Lorenz-Mie theory. We computed scattering properties using the Discrete Dipole Approximation code ADDA for irregularly shaped particles across size parameters x from 0.1 to 65, covering a range in complex refractive index m that includes astrosilicates, pyroxene, enstatite, water-ice, etc. The neural network operates at millisecond timescales while maintaining superior accuracy compared to linear interpolation. Irregular grains exhibit x-dependent deviations from spherical predictions. At small x, cross-sections approach volume-equivalent spheres for low m. At large x, irregular grains show enhanced cross-sections due to greater geometric extension. Increasing m also enhances the absorption cross-section relative to the volume-equivalent spheres. This differential x and m dependence creates mid-IR solid-state features distinct from predictions from spherical grains. Validation against laboratory measurements of forsterite and hematite demonstrates that our neural network captures both qualitative and quantitative behaviors more accurately than spherical models. Millimeter-wavelength applications reveal that spherical grains produce opposite polarization signatures compared to irregular grains, potentially relaxing stringent ~100um grain size constraints in protoplanetary disks. glitterin is publicly available and alleviates the computational barriers to incorporating emission and scattering of realistic grain morphologies. △ Less

Submitted 12 November, 2025; originally announced November 2025.

Comments: 24 pages, 12 figures, accepted for publication in PASP. Training data, trained neural networks, and the python interface are publicly available. Abstract abridged for arXiv requirements

Learning-based Radio Link Failure Prediction Based on Measurement Dataset in Railway Environments

Authors: Po-Heng Chou, Da-Chih Lin, Hung-Yu Wei, Walid Saad, Yu Tsao

Abstract: In this paper, a measurement-driven framework is proposed for early radio link failure (RLF) prediction in 5G non-standalone (NSA) railway environments. Using 10 Hz metro-train traces with serving and neighbor-cell indicators, we benchmark six models, namely CNN, LSTM, XGBoost, Anomaly Transformer, PatchTST, and TimesNet, under varied observation windows and prediction horizons. When the observati… ▽ More In this paper, a measurement-driven framework is proposed for early radio link failure (RLF) prediction in 5G non-standalone (NSA) railway environments. Using 10 Hz metro-train traces with serving and neighbor-cell indicators, we benchmark six models, namely CNN, LSTM, XGBoost, Anomaly Transformer, PatchTST, and TimesNet, under varied observation windows and prediction horizons. When the observation window is three seconds, TimesNet attains the highest F1 score with a three-second prediction horizon, while CNN provides a favorable accuracy-latency tradeoff with a two-second horizon, enabling proactive actions such as redundancy and adaptive handovers. The results indicate that deep temporal models can anticipate reliability degradations several seconds in advance using lightweight features available on commercial devices, offering a practical path to early-warning control in 5G-based railway systems. △ Less

Submitted 13 November, 2025; v1 submitted 11 November, 2025; originally announced November 2025.

Comments: 6 pages, 3 figures, 2 tables, and submitted to IEEE ICC 2026

Beyond English: Toward Inclusive and Scalable Multilingual Machine Translation with LLMs

Authors: Yingfeng Luo, Ziqiang Xu, Yuxuan Ouyang, Murun Yang, Dingyang Lin, Kaiyan Chang, Tong Zheng, Bei Li, Peinan Feng, Quan Du, Tong Xiao, Jingbo Zhu

Abstract: Large language models have significantly advanced Multilingual Machine Translation (MMT), yet the broad language coverage, consistent translation quality, and English-centric bias remain open challenges. To address these challenges, we introduce \textbf{LMT}, a suite of \textbf{L}arge-scale \textbf{M}ultilingual \textbf{T}ranslation models centered on both Chinese and English, covering 60 language… ▽ More Large language models have significantly advanced Multilingual Machine Translation (MMT), yet the broad language coverage, consistent translation quality, and English-centric bias remain open challenges. To address these challenges, we introduce \textbf{LMT}, a suite of \textbf{L}arge-scale \textbf{M}ultilingual \textbf{T}ranslation models centered on both Chinese and English, covering 60 languages and 234 translation directions. During development, we identify a previously overlooked phenomenon of \textbf{directional degeneration}, where symmetric multi-way fine-tuning data overemphasize reverse directions (X $\to$ En/Zh), leading to excessive many-to-one mappings and degraded translation quality. We propose \textbf{Strategic Downsampling}, a simple yet effective method to mitigate this degeneration. In addition, we design \textbf{Parallel Multilingual Prompting (PMP)}, which leverages typologically related auxiliary languages to enhance cross-lingual transfer. Through rigorous data curation and refined adaptation strategies, LMT achieves SOTA performance among models of comparable language coverage, with our 4B model (LMT-60-4B) surpassing the much larger Aya-101-13B and NLLB-54B models by a substantial margin. We release LMT in four sizes (0.6B/1.7B/4B/8B) to catalyze future research and provide strong baselines for inclusive, scalable, and high-quality MMT \footnote{\href{https://github.com/NiuTrans/LMT}{https://github.com/NiuTrans/LMT}}. △ Less

Submitted 10 November, 2025; originally announced November 2025.