Showing 1–50 of 446 results for author: Nair, S

Barium Hexaferrite Thin Films as a Scalable Magnetic-Insulator Platform for Proximity-Engineered Spintronics

Authors: Shyam Sundar Poriah, Sanjana D. S., Agrim Sharma, Sreelakshmi M. Nair, Pankaj Bhardwaj, Laxmipriya Nanda, Aryaman Das, Jagadish Rajendran, R. S. Patel, Manish Jain, Dhavala Suri

Abstract: Rare-earth iron garnets, such as yttrium iron garnet (YIG) and thulium iron garnet (TmIG), are the benchmark magnetic insulators for spintronic and magnonic devices, but achieving usable perpendicular magnetic anisotropy (PMA) in these materials typically relies on substrate strain- engineering, requiring careful lattice-matching and specific growth conditions that constrain ma- terial accessibili… ▽ More Rare-earth iron garnets, such as yttrium iron garnet (YIG) and thulium iron garnet (TmIG), are the benchmark magnetic insulators for spintronic and magnonic devices, but achieving usable perpendicular magnetic anisotropy (PMA) in these materials typically relies on substrate strain- engineering, requiring careful lattice-matching and specific growth conditions that constrain ma- terial accessibility. Here we establish sputter grown barium hexaferrite (BaFe12O19, BaM) as a magnetic-insulator alternative with strong intrinsic perpendicular anisotropy, requiring no strain engineering. X-ray diffraction, transmission electron microscopy and Raman spectroscopy confirm stoichiometric films with atomically smooth surfaces, while first-principles calculations corroborate a robust ferrimagnetic ground state. The films exhibit square out-of-plane hysteresis with a coercive field of nearly 0.1 T. Unlike rare-earth garnets, the perpendicular anisotropy in BaM is intrinsic to its magnetoplumbite crystal structure, arising independent of highly ordered strain. Interfaced with Pt and with exfoliated BiSbTeSe2 (BSTS), BaM induces proximity induced anomalous Hall trans- port, confirming efficient interfacial exchange coupling, while the BSTS/BaM heterostructure shows an additional Hall contribution suggestive of non-collinear interfacial spin textures. These results position BaM thin films as a scalable magnetic-insulator platform for spintronic and topological heterostructure devices beyond the constraints of garnet chemistry. △ Less

Submitted 15 August, 2026; originally announced August 2026.

Set-up and Characterisation of Atmospheric Boundary Layers in the 10'x5' Wind Tunnel

Authors: Sita M. Nair, Kevin Gouder

Abstract: The Atmospheric Boundary Layer (ABL) plays a critical role in influencing objects exposed to atmospheric conditions, making its study crucial. Due to the high cost of real-world testing, this thesis focuses on replicating marine ABLs in a wind tunnel environment. Two profiles were developed: one that served as a framework for establishing commonality among the various international wind engineerin… ▽ More The Atmospheric Boundary Layer (ABL) plays a critical role in influencing objects exposed to atmospheric conditions, making its study crucial. Due to the high cost of real-world testing, this thesis focuses on replicating marine ABLs in a wind tunnel environment. Two profiles were developed: one that served as a framework for establishing commonality among the various international wind engineering standards ('Profile 1'), and a second profile, which is more suitable for modelling the inflow to wind farms in the English Channel and the North Sea ('Profile 2'). The ABLs were generated using Irwin spires without floor roughness elements, and the flow characteristics were measured using Laser Doppler Anemometry (LDA) and a multi-hole probe (MHP). MHP showed a reasonably high accuracy when compared to LDA, with less than 1% deviation in the streamwise velocity and under 5% standard deviation in the streamwise, spanwise, and wall-normal velocity components. The profiles achieved good agreement with target metrics such as normalised velocity and turbulence intensity. Spanwise uniformity and spectral analysis confirmed the robustness of the simulation across varying inflow velocities. Overall, Irwin spires proved to be a cost-effective and reliable method for simulating marine ABLs, offering valuable insights for optimising offshore wind energy systems. △ Less

Submitted 7 August, 2026; originally announced August 2026.

Comments: 65 pages, 82 figures

Decentralized Compute on Untrusted Hardware Using Intel TDX and Encrypted CVMs

Authors: Venish Patidar, Dhruv Bindra, Ahmed Darwich, Josh Brown, Haidong Xia, Sathi Nair

Abstract: The rapid growth of artificial intelligence workloads has generated an unprecedented demand for secure and scalable compute resources. However, centralized cloud providers continue to dominate both pricing and security models. In an increasingly competitive AI landscape, where the compromise of training data or model weights can confer a significant advantage, there is a critical need for a comput… ▽ More The rapid growth of artificial intelligence workloads has generated an unprecedented demand for secure and scalable compute resources. However, centralized cloud providers continue to dominate both pricing and security models. In an increasingly competitive AI landscape, where the compromise of training data or model weights can confer a significant advantage, there is a critical need for a computing infrastructure that safeguards data at rest, in transit, and in use, while remaining affordable and broadly accessible. Furthermore, existing GPU cluster offerings (e.g., 8xH100s, 8xH200s, 8xB200s) create financial barriers that limit access for organizations, startups, and independent researchers seeking secure, high-performance computing environments. This paper introduces a decentralized, confidential computing platform that leverages Intel Trust Domain Extensions (TDX), Intel Trust Authority (ITA) and NVIDIA Confidential Computing (CC) to establish a distributed ecosystem of fully encrypted Confidential Virtual Machines (CVMs). The proposed architecture incentivizes hardware providers to contribute Intel TDX capable compute resources. Each participating provider is provisioned with a freshly instantiated, uniquely encrypted Ubuntu 24.04 CVM, providing data protection across all stages, at rest, in transit, and in use. By decentralizing the confidential computing stack and leveraging confidential computing across independently operated nodes, this work demonstrates a viable alternative to traditional cloud-based infrastructures. The proposed system offers enhanced security assurances, transparent cost structures, and democratized access to enterprise-grade secure compute capabilities, paving the way for a more open, secure, and equitable foundation for next-generation AI development. △ Less

Submitted 23 July, 2026; originally announced July 2026.

Comments: 9 pages, 3 figures. Prior version at Intel Community Blog and manifold.inc

ACM Class: C.2.4; D.4.6

Physics-Based Deep Spatiotemporal Hyperlocal Radar Nowcasting with a Multi-Variable U-Net for High-Resolution Precipitation Forecasting

Authors: Akshay Sunil, Muhammed Rashid, Raja Sekhar Sivaraju, Sushma Nair, Subimal Ghosh

Abstract: Precipitation nowcasting over the immediate 10-90 min period is important for flood management and real-time decision-making in urban regions. Conventional short-range forecasting with high-resolution numerical weather prediction requires frequent data assimilation, model initialization, and spin-up, introducing computational latency. Machine learning provides an alternative by learning storm evol… ▽ More Precipitation nowcasting over the immediate 10-90 min period is important for flood management and real-time decision-making in urban regions. Conventional short-range forecasting with high-resolution numerical weather prediction requires frequent data assimilation, model initialization, and spin-up, introducing computational latency. Machine learning provides an alternative by learning storm evolution directly from high-frequency observations and producing forecasts quickly after training. This is particularly relevant for Mumbai, India, where monsoon convection, land-sea interactions, and localized intense rainfall make short-term prediction difficult. Here, we develop a compact radar-only nowcasting framework that combines multi-elevation reflectivity, Doppler radial velocity, and radial-velocity-gradient proxy features within an encoder-decoder U-Net. Using the most recent radar volume scan, the model predicts 12 future composite reflectivity fields at 7.5-min intervals up to 90 min lead time. The derived velocity magnitude, divergence-like, directional-shear, and vorticity-like channels represent kinematic signatures associated with convergence and boundary interactions without requiring full wind-field retrieval. A high-reflectivity attention module improves sensitivity to convective cores, and physics-guided attribution examines whether the learned sensitivities are meteorologically meaningful. The model is trained using Mumbai Doppler radar observations from May to August 2023 and evaluated on temporally independent events. At 90 min lead time, Critical Success Index values are 0.437, 0.332, and 0.193 for $\geq$10, $\geq$20, and $\geq$30 dBZ thresholds, respectively. Compared with persistence, the model gives lower RMSE and higher spatial correlation at longer lead times. Once trained, it runs on a standard computer, generating nowcasts within seconds for real-time use. △ Less

Submitted 17 July, 2026; originally announced July 2026.

Comments: 10 pages, 5 figures. Submitted to IEEE Transactions on Geoscience and Remote Sensing

Electron Beam Radiolysis-Assisted Growth of Rutile TiO2 Thin Films

Authors: Silu Guo, Nitin Sathish Kumar, Sreejith Nair, Supriya Ghosh, Bharat Jalan, K. Andre Mkhoyan

Abstract: A new approach for growing crystalline thin films is developed that takes advantage of electron beam radiolysis being a constructive force to rearrange atoms into a crystalline structure. It is demonstrated that by irradiating the surface of a TiO2 film by an electron beam supplied by a reflection high energy electron diffraction (RHEED) gun inside the MBE chamber during growth, a crystalline film… ▽ More A new approach for growing crystalline thin films is developed that takes advantage of electron beam radiolysis being a constructive force to rearrange atoms into a crystalline structure. It is demonstrated that by irradiating the surface of a TiO2 film by an electron beam supplied by a reflection high energy electron diffraction (RHEED) gun inside the MBE chamber during growth, a crystalline film can be grown at much lower substrate temperatures, where deposited films typically appear amorphous. Here, rutile TiO2 films were grown using hybrid molecular beam epitaxy (MBE) allowing atomic level control of growth as well as an observation of radiolysis-driven crystallization. Analysis was carried out using a combination of SEM and atomic-resolution STEM imaging. It is also shown that by tuning the temperature of the substrate and the dose of the electron beam, the degree of crystallinity of the film can be controlled. △ Less

Submitted 15 July, 2026; originally announced July 2026.

Room-temperature inversionless diamond nitrogen-vacancy electronic spin maser

Authors: Ali Fawaz, Sarath Raman Nair

Abstract: We propose a method to create a room-temperature maser operating at approximately 2.9~GHz frequency using an ensemble of negatively charged nitrogen-vacancy electronic spins (NV) in diamond, without requiring population inversion. Our method considers a DC magnetic field of a few milli-Tesla (mT) applied along the perpendicular direction of an ensemble of NV spins aligned along a common axis. This… ▽ More We propose a method to create a room-temperature maser operating at approximately 2.9~GHz frequency using an ensemble of negatively charged nitrogen-vacancy electronic spins (NV) in diamond, without requiring population inversion. Our method considers a DC magnetic field of a few milli-Tesla (mT) applied along the perpendicular direction of an ensemble of NV spins aligned along a common axis. This perpendicular magnetic-field creates superposition states of $|m_{\mathrm{s}}=-1\rangle$ and $|m_{\mathrm{s}}=+1\rangle$ of the NV spin's ground state triplet levels and thereby makes it possible to drive all three transitions in the NV spin ground state. We model the system by including optical pumping of the NV spins, near-resonant driving of two transitions, and coupling the third transition to a near-resonant microwave resonator. Numerical estimates using experimentally realizable parameters show that inversionless masing can be achieved inside the microwave resonator using our method. As an application, we show that the output intensity of an inversionless maser ($1.1\times10^{14}$ spins) can be used for magnetic field sensing with a DC sensitivity on the order of a hundred pT/$\sqrt{\mathrm{Hz}}$. Our study opens a new direction in room-temperature diamond NV maser devices for quantum technological applications without the requirement of a strong bias magnetic field, as in conventional NV diamond masers. △ Less

Submitted 8 July, 2026; originally announced July 2026.

Comments: 6 pages, 3 figures, supplementary material (10 pages, 7 figures)

On the deformation theory of chiral quantizations

Authors: Dylan Butson, Sujay Nair

Abstract: We give an operadic approach to deformation quantization of vertex Poisson algebras, a chiral analogue of the traditional problem of deformation quantization of Poisson algebras. Our main result is an order-by-order deformation-obstruction theory for such quantizations, controlled by the chiral analogue of Poisson cohomology. In the special case of chiral quantizations of affine symplectic varieti… ▽ More We give an operadic approach to deformation quantization of vertex Poisson algebras, a chiral analogue of the traditional problem of deformation quantization of Poisson algebras. Our main result is an order-by-order deformation-obstruction theory for such quantizations, controlled by the chiral analogue of Poisson cohomology. In the special case of chiral quantizations of affine symplectic varieties, quantizations of the vertex Poisson algebras of functions on their arc spaces, we prove that this deformation-obstruction theory is controlled by their de Rham cohomology. As another application, we prove that the boundary Virasoro minimal models are rigid under deformations. △ Less

Submitted 25 June, 2026; originally announced June 2026.

Comments: 91 pages, comments welcome

Towards Distributed Inference of LLMs on a P2P Network

Authors: Shabari S Nair, Krishanu Saini

Abstract: Prefix caching can reduce LLM inference latency by reusing KV caches across requests with shared prompts, but cluster-scale reuse is challenging because caches are partitioned across nodes. We propose a decentralized, prefix-cache-aware routing scheme for peer-to-peer LLM serving. Each node maintains a local radix tree of its own cached prefixes and asynchronously refreshed estimates of peer cache… ▽ More Prefix caching can reduce LLM inference latency by reusing KV caches across requests with shared prompts, but cluster-scale reuse is challenging because caches are partitioned across nodes. We propose a decentralized, prefix-cache-aware routing scheme for peer-to-peer LLM serving. Each node maintains a local radix tree of its own cached prefixes and asynchronously refreshed estimates of peer caches using periodic anti-entropy. Requests are routed to the node with the longest estimated prefix match, without centralized coordination or KV-cache transfer. Stale metadata only causes cache misses, not incorrect outputs, making weak consistency sufficient for correctness. Evaluation on simulated MMLU workloads show that decentralized routing improves latency under low communication delay and skewed prefix distributions, while high network latency and affinity-induced hotspots limit its benefits. △ Less

Submitted 7 May, 2026; originally announced June 2026.

On the classification of indecomposable Ekedahl-Oort strata in unitary Shimura varieties, and related Newton polygons

Authors: Emerald Andrews, Deewang Bhamidipati, Maria Fox, Heidi Goodson, Steven R. Groen, Sandra Nair

Abstract: In this paper, we give a complete classification of indecomposable Ekedahl--Oort strata of Shimura varieties associated to the unitary group $\mathsf{GU}(a, b)$ over an odd inert prime. We show that each indecomposable stratum is one of four types: unitary unicycle, unitary bicycle, Serre unicycle, or Serre bicycle; the latter two types are named for a tensor construction of abelian varieties deve… ▽ More In this paper, we give a complete classification of indecomposable Ekedahl--Oort strata of Shimura varieties associated to the unitary group $\mathsf{GU}(a, b)$ over an odd inert prime. We show that each indecomposable stratum is one of four types: unitary unicycle, unitary bicycle, Serre unicycle, or Serre bicycle; the latter two types are named for a tensor construction of abelian varieties developed by Serre. We provide an algorithm that translates the description of a stratum in terms of words in the alphabet $\{\texttt{f},\texttt{v}\}$ to the corresponding Weyl group coset representative. Finally, using a $p$-adic lift, we construct a `tautological' point in each Ekedahl--Oort stratum, and compute its Newton polygon. As an application, we show that the indecomposable Ekedahl--Oort strata corresponding to unitary unicycles and Serre unicycles always intersect the supersingular locus. △ Less

Submitted 23 July, 2026; v1 submitted 15 June, 2026; originally announced June 2026.

Comments: 41 pages, 7 figures, 2 tables. Updates in version 2: Added Lemma 4.27 determining objectivity of the Serre tensor construction, and an appendix on counting the number of unitary BT1 modules of given indecomposable type for a fixed signature. Comments are welcome!

Report number: MPIM-Bonn-2026 MSC Class: 14G35; 14G17; 11G10; 14K10

Tidal Stripping of Matter Bound to the Secondary in Extreme Mass-Ratio Inspirals

Authors: Sreejith Nair, Sayak Datta

Abstract: Environmental studies of extreme mass-ratio inspirals (EMRIs) have focused almost entirely on matter surrounding the primary supermassive black hole. We instead consider matter bound to the stellar-mass secondary (e.g., gas or dark matter); which can be progressively tidally stripped during the LISA-band inspiral. This changes the bound mass of the inspiraling object, modifying the gravitational-w… ▽ More Environmental studies of extreme mass-ratio inspirals (EMRIs) have focused almost entirely on matter surrounding the primary supermassive black hole. We instead consider matter bound to the stellar-mass secondary (e.g., gas or dark matter); which can be progressively tidally stripped during the LISA-band inspiral. This changes the bound mass of the inspiraling object, modifying the gravitational-wave (GW) phase at leading order in the secondary mass. Furthermore, as the signal interpolates from an initially dressed inspiral to a nearly bare one, it can produce a characteristic inflection in the residual phase with constant mass waveform templates. Even for an environmental mass $\sim 10^{-3}\,M_{\odot}$, the cumulative dephasing relative to in band initial bound mass waveform can be larger than unity. In subsolar mass cases, the relative dephasing can reach $O(10^3)\, \rm rad$. Neglecting this effect may bias inferred EMRI parameters at the level of the fractional change in the in-band bound mass. The tidal stripping phenomena carry information about the mass and the compactness of the bound matter, enabling probes of sub-AU, planetary- to subsolar-mass environments surrounding stellar-mass black holes. △ Less

Submitted 12 June, 2026; originally announced June 2026.

Comments: 8 pages, 4 figures

Nuclear matter and proton parton distributions in a light-front Hamiltonian framework

Authors: Xiaoyi Wu, Sreeraj Nair, Satvir Kaur, Chandan Mondal, Jiangshan Lan, Xingbo Zhao, J. P. B. C. de Melo, Tobias Frederico

Abstract: We develop a light-front Hamiltonian formulation of symmetric nuclear matter within the quark-meson coupling model, using Basis Light-Front Quantization to solve the in-medium nucleon eigenvalue problem. The Hamiltonian incorporates confinement in the valence sector and is truncated to include up to one dynamical gluon. Medium effects are introduced via scalar and vector mean fields, yielding a se… ▽ More We develop a light-front Hamiltonian formulation of symmetric nuclear matter within the quark-meson coupling model, using Basis Light-Front Quantization to solve the in-medium nucleon eigenvalue problem. The Hamiltonian incorporates confinement in the valence sector and is truncated to include up to one dynamical gluon. Medium effects are introduced via scalar and vector mean fields, yielding a self-consistent, density-dependent effective quark mass and modified nucleon structure. The resulting energy per nucleon, pressure, and incompressibility are consistent with empirical constraints at the saturation point. At nuclear saturation density, the gluon probability in the nucleon wave function increases slightly, while the valence probability and quark momentum fraction decrease. The unpolarized quark and gluon distributions show a noticeable enhancement at large momentum fraction ($x \gtrsim 0.4$), illustrated at an evolved scale of $Q^{2} = 10 \mathrm{GeV}^{2}$. △ Less

Submitted 29 May, 2026; originally announced May 2026.

Comments: 10 pages, 4 figures

Magneto-Optical Detection of Anisotropic Spin Currents in Altermagnetic RuO2

Authors: Joongwon Lee, Jeonglyul Kim, Sreejith Nair, Seung Gyo Jeong, Changi Kim, Jae-Pil So, Bohm-Jung Yang, Bharat Jalan, Hyobin Yoo, Farhan Rana, Taekoo Oh, Hong-Gyu Park

Abstract: Altermagnets are a recently identified class of collinear antiferromagnets that host large spin-split electronic bands, offering a promising platform for efficient spin-current generation. Among proposed candidates, the metallic oxide RuO2 is predicted to exhibit strong altermagnetic spin splitting; however, whether it sustains robust magnetic order beyond the ultrathin thickness limit remains unr… ▽ More Altermagnets are a recently identified class of collinear antiferromagnets that host large spin-split electronic bands, offering a promising platform for efficient spin-current generation. Among proposed candidates, the metallic oxide RuO2 is predicted to exhibit strong altermagnetic spin splitting; however, whether it sustains robust magnetic order beyond the ultrathin thickness limit remains unresolved. Here, we employ optical probes to investigate charge-to-spin conversion in a 12-nm-thick (101)-oriented RuO2 film grown on sapphire. Polarization-resolved second-harmonic generation reveals nonlinear optical responses consistent with the surface symmetry and Néel order of RuO2. Under an applied current, both second-harmonic generation and polar magneto-optical Kerr effect measurements detect a pronounced, directionally anisotropic spin polarization, exhibiting enhanced signals for current along [010] and strongly suppressed responses for current along [-101], in agreement with the symmetry of the altermagnetic spin-splitter effect. Non-magnetic or Rashba-type mechanisms cannot explain this symmetry-selective response. Scanning transmission electron microscopy further reveals that substantial strain persists even in relatively thick films, providing a possible explanation for the observed behavior. Therefore, these results establish RuO2 as an efficient spin source and demonstrate the potential of altermagnets for field-free spintronic devices. △ Less

Submitted 26 May, 2026; originally announced May 2026.

A Reinforcement Learning Inspired Latent Yield Based Adaptive Algorithm Switching Mechanism

Authors: Jayprakash S. Nair, Jimson Mathew, Shivashankar B. Nair

Abstract: Selecting the most suitable algorithm for a given problem instance remains a challenging task, particularly in online or dynamic environments where problem characteristics evolve over time. Relying solely on instantaneous performance metrics can result in a reactive and unstable behaviour, often leading to suboptimal algorithm switching. This paper introduces a computationally efficient approach f… ▽ More Selecting the most suitable algorithm for a given problem instance remains a challenging task, particularly in online or dynamic environments where problem characteristics evolve over time. Relying solely on instantaneous performance metrics can result in a reactive and unstable behaviour, often leading to suboptimal algorithm switching. This paper introduces a computationally efficient approach for aggregating an algorithm's performance across multiple problem instances that is fairly immune to erratic variations in instance features. Inspired by features inherent to Reinforcement Learning (RL), this technique encapsulates rewards and penalties into a latent yield that, in turn, triggers exploitation and exploration, consequently resulting in adaptive algorithm switching. The proposed technique employs island models, inspired by Genetic Algorithms, to facilitate parallel exploration and performance exchanges among algorithm populations inhabiting local repertoires. Experimental evaluations on sorting algorithms and robotic obstacle avoidance tasks demonstrate the feasibility and effectiveness of the approach, highlighting its potential in domains where adaptive algorithm selection is critical. △ Less

Submitted 23 May, 2026; originally announced May 2026.

Comments: Accepted and published in the Proceedings of the 29th European Conference on Applications of Evolutionary Computation (EvoApplications 2026), held as part of EvoStar 2026, Toulouse, France, April 8 to 10, 2026. Lecture Notes in Computer Science (LNCS), Springer Nature Switzerland

Journal ref: Applications of Evolutionary Computation, EvoApplications 2026, LNCS, Springer Nature Switzerland, 2026

On the Harris-Viehmann conjecture for Hodge-Newton reducible local Shimura data of abelian type

Authors: Sandra Nair, Xinyu Zhou

Abstract: We address a new case of the Harris-Viehmann conjecture, which establishes a parabolic induction formula on the cohomology groups associated to non-basic local Shimura data. It follows that all supercuspidal representations on a Shimura variety are concentrated along the basic locus, making the conjecture relevant to the Langlands program. Historically, many cases of the Harris-Viehmann conjecture… ▽ More We address a new case of the Harris-Viehmann conjecture, which establishes a parabolic induction formula on the cohomology groups associated to non-basic local Shimura data. It follows that all supercuspidal representations on a Shimura variety are concentrated along the basic locus, making the conjecture relevant to the Langlands program. Historically, many cases of the Harris-Viehmann conjecture have been approached with the additional condition of Hodge-Newton reducibility on the underlying local Shimura datum. Building on previous work by Mantovan (EL/PEL case) and Hong (Hodge case), we extend the proof of the conjecture to unramified non-basic local Shimura data of abelian type under the assumption of Hodge-Newton reducibility. We leverage Shen's construction of Rapoport-Zink spaces of abelian type at the hyperspecial level. △ Less

Submitted 29 May, 2026; v1 submitted 22 May, 2026; originally announced May 2026.

Comments: TeX problems fixed. 31 pages of content. Comments are welcome!

PMF-CL: Pareto-Minimal-Forgetting Continual Learner for Conflicting Tasks

Authors: Srijith Nair, Atilla Eryilmaz, Jia Liu

Abstract: In the literature, many continual learning (CL) algorithms have been proposed to address the issue of catastrophic forgetting in ML models (i.e., learning new tasks leads to the loss of performance on previously learned tasks). Although all CL approaches use some form of memory to retain information about past tasks, a grounded understanding of what information needs to be stored to minimize catas… ▽ More In the literature, many continual learning (CL) algorithms have been proposed to address the issue of catastrophic forgetting in ML models (i.e., learning new tasks leads to the loss of performance on previously learned tasks). Although all CL approaches use some form of memory to retain information about past tasks, a grounded understanding of what information needs to be stored to minimize catastrophic forgetting remains elusive. Recently, it has been recognized that under the strong assumption of the existence of a common global minimizer over all tasks, catastrophic forgetting can be completely avoided. However, in practice, tasks rarely have a common global minimizer, and a certain amount of forgetting is inevitable. In this paper, we propose a foundational framework for principled and systematic CL of conflicting tasks using a multi-task learning (MTL) perspective. The approach is based on finding Pareto-optimal solutions, i.e., the solutions which, by definition, minimally forget the previous tasks in the Pareto sense. We derive Pareto-minimal-forgetting CL algorithms for linear and basis-function regression, and general loss functions which have a quadratic upper bound, e.g., logistic regression. For quadratic problems, PMF-CL uses memory-efficient iterative updates with a static memory footage of $\mathcal{O}(d^2)$ for models with $d$ parameters. △ Less

Submitted 28 May, 2026; v1 submitted 18 May, 2026; originally announced May 2026.

Comments: 25 pages, 4 figures, 4 algorithms

Gravitational form factors of light mesons from Basis Light-Front Quantization

Authors: Amrita Sain, Sreeraj Nair, Chandan Mondal, Xingbo Zhao, James P. Vary

Abstract: We compute the gravitational form factors (GFFs) of the pion and kaon using their light-front wave functions within the Basis Light-Front Quantization framework. The wave functions are obtained by solving a light-front effective Hamiltonian that incorporates three-dimensional confinement along with a color-singlet Nambu--Jona-Lasinio interaction between the constituent quark and antiquark. The for… ▽ More We compute the gravitational form factors (GFFs) of the pion and kaon using their light-front wave functions within the Basis Light-Front Quantization framework. The wave functions are obtained by solving a light-front effective Hamiltonian that incorporates three-dimensional confinement along with a color-singlet Nambu--Jona-Lasinio interaction between the constituent quark and antiquark. The form factor $A(Q^2)$ is found to be in overall agreement with recent lattice QCD and dispersive results. In contrast, $D(Q^2)$ is enhanced in magnitude at low $Q^2$ relative to both lattice QCD and dispersive determinations. This behavior arises from extracting the $D$-term using transverse components of the QCD energy--momentum tensor, which are more sensitive to the small-$x$ region and to light-front zero-mode effects in the present truncated framework. Using the resulting GFFs, we determine the mass (matter) and mechanical radii of the pion and kaon and analyze their mechanical structure through the corresponding pressure and shear-force distributions. △ Less

Submitted 18 May, 2026; originally announced May 2026.

Comments: 11 pages, 5 figures

Do Enterprise Systems Need Learned World Models? The Importance of Context to Infer Dynamics

Authors: Jishnu Sethumadhavan Nair, Patrice Bechard, Rishabh Maheshwary, Surajit Dasgupta, Sravan Ramachandran, Aakash Bhagat, Shruthan Radhakrishna, Pulkit Pattnaik, Johan Obando-Ceron, Shiva Krishna Reddy Malay, Sagar Davasam, Seganrasan Subramanian, Vipul Mittal, Sridhar Krishna Nemala, Christopher Pal, Srinivas Sunkara, Sai Rajeswar

Abstract: World models enable agents to anticipate the effects of their actions by internalizing environment dynamics. In enterprise systems, however, these dynamics are often defined by tenant-specific business logic that varies across deployments and evolves over time, making models trained on historical transitions brittle under deployment shift. We ask a question the world-models literature has not addr… ▽ More World models enable agents to anticipate the effects of their actions by internalizing environment dynamics. In enterprise systems, however, these dynamics are often defined by tenant-specific business logic that varies across deployments and evolves over time, making models trained on historical transitions brittle under deployment shift. We ask a question the world-models literature has not addressed: when the rules can be read at inference time, does an agent still need to learn them? We argue, and demonstrate empirically, that in settings where transition dynamics are configurable and readable, runtime discovery complements offline training by grounding predictions in the active system instance. We propose enterprise discovery agents, which recover relevant transition dynamics at runtime by reading the system's configuration rather than relying solely on internalized representations. We introduce CascadeBench, a reasoning-focused benchmark for enterprise cascade prediction that adopts the evaluation methodology of World of Workflows on diverse synthetic environments, and use it together with deployment-shift evaluation to show that offline-trained world models can perform well in-distribution but degrade as dynamics change, whereas discovery-based agents are more robust under shift by grounding their predictions in the current instance. Our findings suggest that, in configurable enterprise environments, agents should not rely solely on fixed internalized dynamics, but should incorporate mechanisms for discovering relevant transition logic at runtime. △ Less

Submitted 12 May, 2026; originally announced May 2026.

SHIELD: Scalable Optimal Control with Certification using Duality and Convexity

Authors: Hansung Kim, Siddharth H. Nair, Francesco Borrelli

Abstract: We present SHIELD, a hierarchical algorithm that reduces both the decision-variable dimension and the constraint set in $\ell_1$-regularized convex programs. From strong convexity and Lagrangian duality, we derive certificates that \emph{safely} discard constraints and decision variables while guaranteeing that all removed constraints remain satisfied and all removed variables are null. To further… ▽ More We present SHIELD, a hierarchical algorithm that reduces both the decision-variable dimension and the constraint set in $\ell_1$-regularized convex programs. From strong convexity and Lagrangian duality, we derive certificates that \emph{safely} discard constraints and decision variables while guaranteeing that all removed constraints remain satisfied and all removed variables are null. To further accelerate the proposed algorithm, we propose a transformer-based deep neural network to guide the dual certificate inference. We validate SHIELD on stochastic model predictive control (SMPC) in complex, multi-modal traffic scenarios, comparing against a full-dimensional SMPC policy. Numerical simulations demonstrate order-of-magnitude computational speedups while preserving feasibility and closed-loop safety, highlighting the practicality of certifiably safe, lightweight MPC in complex driving scenes. △ Less

Submitted 11 May, 2026; v1 submitted 9 May, 2026; originally announced May 2026.

Baryon Bethe-Salpeter Equation in Minkowski-Space QCD$_2$

Authors: Satvir Kaur, Sreeraj Nair, Chandan Mondal, Jiangshan Lan, Xingbo Zhao, J. P. B. C. de Melo, Tobias Frederico

Abstract: We study the three-quark ladder Bethe--Salpeter equation in Minkowski-space QCD$_2$ in the light-cone gauge. Using the quasi-potential expansion, we project the baryon equation onto the light front and show that, at leading order in the valence truncation, the resulting mass-squared eigenvalue equation is equivalent to the Bars--Durgut equation. We also derive the endpoint power-law behavior of th… ▽ More We study the three-quark ladder Bethe--Salpeter equation in Minkowski-space QCD$_2$ in the light-cone gauge. Using the quasi-potential expansion, we project the baryon equation onto the light front and show that, at leading order in the valence truncation, the resulting mass-squared eigenvalue equation is equivalent to the Bars--Durgut equation. We also derive the endpoint power-law behavior of the valence wave function in terms of the quark mass and coupling, closely paralleling the original 't Hooft analysis for mesons. The resulting three-quark equation is solved numerically for $N_c=3$, and the ground-state baryon mass is found to be in reasonable agreement with previous light-cone quantization results in QCD$_2$, suggesting that the valence sector provides the dominant contribution to the ground state. The excited-state spectrum further yields a Regge trajectory that captures the overall trend of the experimental nucleon spectrum, and we compute selected structure observables, including parton distribution functions, double distribution amplitudes, and coordinate-space densities. This framework provides a useful confining test bed for Minkowski-space bound-state methods and for future developments toward confining formulations in 3+1 dimensions beyond the valence truncation. △ Less

Submitted 7 May, 2026; originally announced May 2026.

Comments: 11 pages, 7 figures

The quotient problem for linear recurrence sequences

Authors: Parvathi S Nair, S. S. Rout

Abstract: Let $\{U(m)\}_{m\in \N}$ and $\{V(n)\}_{n\in \N}$ be linear recurrence sequences. It is a well-known Diophantine problem to determine the finiteness of the set of natural numbers $n$ such that the ratio $U(n)/V(n)$ is an integer. We study the finiteness problem for the set $(m, n)\in \mathbb{N}^2$ such that there exist non-zero positive integers $d_{m, n}$ satisfying $\log |d_{m, n}|=o(n)$, and… ▽ More Let $\{U(m)\}_{m\in \N}$ and $\{V(n)\}_{n\in \N}$ be linear recurrence sequences. It is a well-known Diophantine problem to determine the finiteness of the set of natural numbers $n$ such that the ratio $U(n)/V(n)$ is an integer. We study the finiteness problem for the set $(m, n)\in \mathbb{N}^2$ such that there exist non-zero positive integers $d_{m, n}$ satisfying $\log |d_{m, n}|=o(n)$, and $d_{m, n}U(m)/V(n)$ is an element from a finitely generated subring of $\C$. In particular, we prove that for $m\neq n $, there exists a polynomial $P$ such that $d_{m, n}P(n)U(m)/V(n)$ is a multi-recurrence and $V(n)/P(n)$ is a linear recurrence and for $m=n$ both $d_{m, n}P(n)U(m)/V(n)$ and $V(n)/P(n)$ are linear recurrences. To prove our results, we employ Schmidt's subspace theorem, and the concept of moving hyperplanes, moving polynomials, and moving points. △ Less

Submitted 7 May, 2026; originally announced May 2026.

MSC Class: Primary 11J87; Secondary 11B37; 11J25

What Matters in Practical Learned Image Compression

Authors: Kedar Tatwawadi, Parisa Rahimzadeh, Zhanghao Sun, Zhiqi Chen, Ziyun Yang, Sanjay Nair, Divija Hasteer, Oren Rippel

Abstract: One of the major differentiators unlocked by learned codecs relative to their hard-coded traditional counterparts is their ability to be optimized directly to appeal to the human visual system. Despite this potential, a perceptual yet practical image codec is yet to be proposed. In this work, we aim to close this gap. We conduct a comprehensive study of the key modeling choices that govern the d… ▽ More One of the major differentiators unlocked by learned codecs relative to their hard-coded traditional counterparts is their ability to be optimized directly to appeal to the human visual system. Despite this potential, a perceptual yet practical image codec is yet to be proposed. In this work, we aim to close this gap. We conduct a comprehensive study of the key modeling choices that govern the design of a practical learned image codec, jointly optimized for perceptual quality and runtime -- including within the ablations several novel techniques. We then perform performance-aware neural architecture search over millions of backbone configurations to identify models that achieve the target on-device runtime while maximizing compression performance as captured by perceptual metrics. We combine the various optimizations to construct a new codec that achieves a significantly improved tradeoff between speed and perceptual quality. Based on rigorous subjective user studies, it provides 2.3-3x bitrate savings against AV1, AV2, VVC, ECM and JPEG-AI, and 20-40% bitrate savings against the best learned codec alternatives. At the same time, on an iPhone 17 Pro Max, it encodes 12MP images as fast as 230ms, and decodes them in 150ms -- faster than most top ML-based codecs run on a V100 GPU. △ Less

Submitted 6 May, 2026; originally announced May 2026.

LA-Pose: Latent Action Pretraining Meets Pose Estimation

Authors: Zhengqing Wang, Saurabh Nair, Prajwal Chidananda, Pujith Kachana, Samuel Li, Matthew Brown, Yasutaka Furukawa

Abstract: This paper revisits camera pose estimation through the lens of self-supervised pretraining, focusing on inverse-dynamics pretraining as a scalable alternative to the current trend of fully supervised training with 3D annotations. Concretely, we employ inverse- and forward-dynamics models to learn latent action representations, similar to Genie from large-scale driving videos. Our idea is simple ye… ▽ More This paper revisits camera pose estimation through the lens of self-supervised pretraining, focusing on inverse-dynamics pretraining as a scalable alternative to the current trend of fully supervised training with 3D annotations. Concretely, we employ inverse- and forward-dynamics models to learn latent action representations, similar to Genie from large-scale driving videos. Our idea is simple yet effective. Existing methods use latent actions in their original capacity, that is, as action conditioning of world-models or as proxies of robot action parameters in policy networks. Our method, dubbed LA-Pose, repurposes the latent action features as inputs to a camera pose estimator, finetuned on a limited set of high-quality 3D annotations. This formulation enables accurate and generalizable pose prediction while maintaining feed-forward efficiency. Extensive experiments on driving benchmarks show that LA-Pose achieves competitive and even superior performance to state-of-the-art methods while using orders of magnitude less labeled data. Concretely, on the Waymo and PandaSet benchmarks, LA-Pose achieves over 10% higher pose accuracy than recent feed-forward methods. To our knowledge, this work is the first to demonstrate the power of inverse-dynamics self-supervised learning for pose estimation. △ Less

Submitted 30 April, 2026; originally announced April 2026.

Comments: Project page: https://la-pose.github.io/

Local tensor-train surrogates for quantum learning models

Authors: Sreeraj Rajindran Nair, Christopher Ferrie

Abstract: A key bottleneck in quantum machine learning is the computational cost of repeated quantum circuit evaluations during the inference phase. To address this, we present a framework for constructing fast, cheap, provably accurate classical tensor-train surrogates of fully trained quantum machine learning models within local patches of their input data space. The approach combines Taylor polynomial ap… ▽ More A key bottleneck in quantum machine learning is the computational cost of repeated quantum circuit evaluations during the inference phase. To address this, we present a framework for constructing fast, cheap, provably accurate classical tensor-train surrogates of fully trained quantum machine learning models within local patches of their input data space. The approach combines Taylor polynomial approximation with a tensor-train (TT) representation and embeds it in a statistical learning paradigm via empirical risk minimization. In our analysis, the Taylor-TT construction serves as a deterministic error certificate proving that the TT hypothesis class contains a good approximation; empirical risk minimization then provably recovers a surrogate with controlled generalization error and explicit bounds. This translates into three independently controllable error sources: (i) Taylor truncation error controlled by the patch radius $r$ and polynomial degree $p$, (ii) TT approximation error controlled by the bond dimension $χ$, and (iii) statistical estimation error. While the parameter count scales polynomially in the number of data dimensions $N$, i.e., $d_{\mathrm{eff}} = N(p+1)χ^2$ rather than the naive $(p+1)^N$, the worst-case constants inherit an exponential factor through the tensor-product feature norm during Taylor polynomial embedding onto TT. This cleanly separates representation complexity from feature-induced constants. Our risk bounds and sample complexity depend explicitly on the local patch radius $r$. △ Less

Submitted 28 April, 2026; originally announced April 2026.

Comments: 26 pages

Computing the Exchange Number in Graphs with respect to Cycle Convexity

Authors: Revathy S. Nair, Bijo S. Anand, Julliano R. Nascimento

Abstract: Given a graph $G$, a subset $S \subseteq V(G)$ is \textit{cycle convex}, if for any vertex $v \in V(G) \setminus S$, the induced subgraph, $G[S \cup \{v\}]$ cannot form a cycle containing the vertex $v$. The \textit{exchange number} of $G$, denoted by $e_{cc}(G)$ is the maximum cardinality of an $\textit{$E$-independent}$ set of $G$. This paper studies the computational complexity of determining t… ▽ More Given a graph $G$, a subset $S \subseteq V(G)$ is \textit{cycle convex}, if for any vertex $v \in V(G) \setminus S$, the induced subgraph, $G[S \cup \{v\}]$ cannot form a cycle containing the vertex $v$. The \textit{exchange number} of $G$, denoted by $e_{cc}(G)$ is the maximum cardinality of an $\textit{$E$-independent}$ set of $G$. This paper studies the computational complexity of determining the exchange number of graphs and provides exact values for some graph classes. Given a graph $G$ and a positive integer $k$, we show that deciding whether $e_{cc}(G) \geq k$ is NP-complete even if $G$ is a $K_5$-free graph. In contrast, we characterize all $n$-vertex graphs $G$ with exchange number $n-1$ and obtain closed formulas for chordal graphs $G$ whose blocks lie in a single chain, which leads to polynomial-time algorithms for computing $e_{cc}(G)$. We also establish a lower bound for the exchange number of the Cartesian product of general graphs and by using the results of Anand et al. \cite{bijo2}, we derive an explicit formula for the exchange number of strong and lexicographic graph products. △ Less

Submitted 22 April, 2026; originally announced April 2026.

MSC Class: 05C69; 05C76; 05C85

Carathéodory Number in Cycle Convexity

Authors: Revathy S. Nair, Bijo S. Anand, Ullas Chandran S. V., Julliano R. Nascimento

Abstract: Let $G$ be a graph and $S \subseteq V(G)$. In the cycle convexity, we say that $S$ is \textit{cycle convex} if for any $u\in V(G)\setminus S$, the induced subgraph of $S\cup\{u\}$ contains no cycle that includes $u$. The \textit{cycle convex hull} of $S$, denoted by $\hullc (S)$, is the smallest cycle convex set containing $S$. A set $S \subseteq V(G)$ is said to be \textit{Carathéodory independen… ▽ More Let $G$ be a graph and $S \subseteq V(G)$. In the cycle convexity, we say that $S$ is \textit{cycle convex} if for any $u\in V(G)\setminus S$, the induced subgraph of $S\cup\{u\}$ contains no cycle that includes $u$. The \textit{cycle convex hull} of $S$, denoted by $\hullc (S)$, is the smallest cycle convex set containing $S$. A set $S \subseteq V(G)$ is said to be \textit{Carathéodory independent} if there exists a vertex $u \in \hullc(S) $ such that $u \notin\displaystyle \bigcup_{a \in S} \hullc (S \setminus \{a\}) $, and the Carathéodory number $\car(G)$ is the maximum size of such a set. In this paper, we prove that given a graph $G$ and $k \in \mathbb{N}$, deciding whether $\car(G) \geq k$ is \NP-complete, even when $G$ is bipartite. On the other hand, we derive exact values and constant upper bounds for several graph classes, leading to polynomial-time algorithms. Some of them include forests, cycles, complete graphs, complete multipartite, split, and $P_4$-sparse graphs. In addition, we present a characterization of $n$-vertex graphs $G$ with extremal values near to $n$, including $\car(G) = n-1$ and $\car(G) = n-2$. Furthermore, we investigate the behavior of the Carathéodory number under graph products such as the strong, lexicographic, and Cartesian products. △ Less

Submitted 21 April, 2026; originally announced April 2026.

$π_{0.7}$: a Steerable Generalist Robotic Foundation Model with Emergent Capabilities

Authors: Physical Intelligence, Bo Ai, Ali Amin, Raichelle Aniceto, Ashwin Balakrishna, Greg Balke, Kevin Black, George Bokinsky, Shihao Cao, Thomas Charbonnier, Vedant Choudhary, Foster Collins, Ken Conley, Grace Connors, James Darpinian, Karan Dhabalia, Maitrayee Dhaka, Jared DiCarlo, Danny Driess, Michael Equi, Adnan Esmail, Yunhao Fang, Chelsea Finn, Catherine Glossop, Thomas Godden , et al. (63 additional authors not shown)

Abstract: We present a new robotic foundation model, called $π_{0.7}$, that can enable strong out-of-the-box performance in a wide range of scenarios. $π_{0.7}$ can follow diverse language instructions in unseen environments, including multi-stage tasks with various kitchen appliances, provide zero-shot cross-embodiment generalization, for example enabling a robot to fold laundry without seeing the task bef… ▽ More We present a new robotic foundation model, called $π_{0.7}$, that can enable strong out-of-the-box performance in a wide range of scenarios. $π_{0.7}$ can follow diverse language instructions in unseen environments, including multi-stage tasks with various kitchen appliances, provide zero-shot cross-embodiment generalization, for example enabling a robot to fold laundry without seeing the task before, and perform challenging tasks such as operating an espresso machine out of the box at a level of performance that matches much more specialized RL-finetuned models. The main idea behind $π_{0.7}$ is to use diverse context conditioning during training. This conditioning information, contained in the prompt, makes it possible to steer the model precisely to perform many tasks with different strategies. It is conditioned not just on a language command that describes what it should do, but on additional multimodal information that also describes the manner or strategy in which it should do it, including metadata about task performance and subgoal images. This enables $π_{0.7}$ to use very diverse data, including demonstrations, potentially suboptimal (autonomous) data including failures, and data from non-robot sources. Our experiments evaluate $π_{0.7}$ across numerous tasks with multiple robot platforms, on tasks that require speed and dexterity, language following, and compositional task generalization. △ Less

Submitted 24 April, 2026; v1 submitted 16 April, 2026; originally announced April 2026.

Comments: Website: https://www.pi.website/blog/pi07

Filling in the Mechanisms: How do LMs Learn Filler-Gap Dependencies under Developmental Constraints?

Authors: Atrey Desai, Sathvik Nair

Abstract: For humans, filler-gap dependencies require a shared representation across different syntactic constructions. Although causal analyses suggest this may also be true for LLMs (Boguraev et al., 2025), it is still unclear if such a representation also exists for language models trained on developmentally feasible quantities of data. We applied Distributed Alignment Search (DAS, Geiger et al. (2024))… ▽ More For humans, filler-gap dependencies require a shared representation across different syntactic constructions. Although causal analyses suggest this may also be true for LLMs (Boguraev et al., 2025), it is still unclear if such a representation also exists for language models trained on developmentally feasible quantities of data. We applied Distributed Alignment Search (DAS, Geiger et al. (2024)) to LMs trained on varying amounts of data from the BabyLM challenge (Warstadt et al., 2023), to evaluate whether representations of filler-gap dependencies transfer between wh-questions and topicalization, which greatly vary in terms of their input frequency. Our results suggest shared, yet item-sensitive mechanisms may develop with limited training data. More importantly, LMs still require far more data than humans to learn comparable generalizations, highlighting the need for language-specific biases in models of language acquisition. △ Less

Submitted 15 April, 2026; originally announced April 2026.

Comments: To be published in the 64th Annual Meeting of the Association for Computational Linguistics

Across the Levels of Analysis: Explaining Predictive Processing in Humans Requires More Than Machine-Estimated Probabilities

Authors: Sathvik Nair, Colin Phillips

Abstract: Under the lens of Marr's levels of analysis, we critique and extend two claims about language models (LMs) and language processing: first, that predicting upcoming linguistic information based on context is central to language processing, and second, that many advances in psycholinguistics would be impossible without large language models (LLMs). We further outline future directions that combine t… ▽ More Under the lens of Marr's levels of analysis, we critique and extend two claims about language models (LMs) and language processing: first, that predicting upcoming linguistic information based on context is central to language processing, and second, that many advances in psycholinguistics would be impossible without large language models (LLMs). We further outline future directions that combine the strengths of LLMs with psycholinguistic models. △ Less

Submitted 10 April, 2026; originally announced April 2026.

Comments: 9 pages, Behavioral & Brain Sciences Commentary on Futrell & Mahowald (forthcoming)

The Quantum Education Ecosystem: A Review of Global Initiatives, Methods, and Challenges

Authors: Sara Ayman Metwalli, Aryan Iliat, Steven Thomas, Suresh Nair, Zizwe A. Chase, Russell R. Ceballos

Abstract: Quantum information science and engineering (QISE) is advancing rapidly, creating an urgent demand for a quantum-literate, technically proficient workforce. Despite this need, quantum education initiatives remain fragmented across regions, educational levels, and instructional approaches, which constrains their scalability and overall impact. This paper offers a structured analysis of the current… ▽ More Quantum information science and engineering (QISE) is advancing rapidly, creating an urgent demand for a quantum-literate, technically proficient workforce. Despite this need, quantum education initiatives remain fragmented across regions, educational levels, and instructional approaches, which constrains their scalability and overall impact. This paper offers a structured analysis of the current quantum education ecosystem by synthesizing global initiatives, pedagogical strategies, and emerging trends. Quantum education is examined through a dual framework that considers both learner progression and instructional methodology, emphasizing the evolution of educational approaches from conceptual exposure to formal reasoning and practical application. Analysis of data from international programs and academic literature reveals key challenges, including inequitable access, absence of standardized curricula, limited empirical evaluation, and discontinuities between educational stages. Quantum education is more accurately conceptualized as a non-linear ecosystem rather than a traditional pipeline, characterized by multiple entry points, feedback mechanisms, and critical transition gaps. Based on this perspective, directions are proposed for developing more coherent, inclusive, and scalable educational frameworks that align with workforce requirements and technological progress. This work presents a unified perspective on the quantum education landscape and outlines actionable strategies to enhance global quantum literacy and workforce preparedness. △ Less

Submitted 7 April, 2026; originally announced April 2026.

A Computational Framework for Cross-Domain Mission Design and Onboard Cognitive Decision Support

Authors: J. de Curtò, Adrianne Schneider, Ricardo Yanez, María Begara, Álvaro Rodríguez, Javier López, Martina Fraga, Ignacio Gómez, Arman Akdag, Sumit Kulkarni, Siddhant Nair, Kiyan Govender, Eian Wratchford, Eli Lynskey, Seamus Dunlap, Cooper Nervick, Nicolas Tête, Rocío Fernández, Pablo González, Elena Municio, I. de Zarzà

Abstract: The design of distributed autonomous systems for operation beyond reliable ground contact presents a fundamental tension: as round-trip communication latency grows, the set of decisions delegable to ground operators shrinks. This paper establishes a unified computational methodology for quantifying and comparing this constraint across seven heterogeneous mission architectures, spanning Earth low-o… ▽ More The design of distributed autonomous systems for operation beyond reliable ground contact presents a fundamental tension: as round-trip communication latency grows, the set of decisions delegable to ground operators shrinks. This paper establishes a unified computational methodology for quantifying and comparing this constraint across seven heterogeneous mission architectures, spanning Earth low-orbit surveillance constellations, Mars orbital navigation systems, autonomous underwater mine-clearing swarms, deep-space inter-satellite link networks, and outer-planet in-situ buoy platforms. We introduce the Autonomy Necessity Score, a log-domain latency metric mapping each system continuously from the ground-dependent to the fully-autonomous regime, grounded in nine independently validated computational studies covering Walker spherical-cap coverage mechanics, infrared Neyman-Pearson detection, Extended Kalman Filter hypersonic tracking, cross-mission RF and acoustic link budgets spanning seven orders of magnitude in range, Monte Carlo science-yield sensitivity for TDMA inter-satellite protocols, cross-architecture power budget sizing, distributed magnetic-signature formation emulation, and Arrhenius-corrected cryogenic swarm reliability. Building on this foundation, we evaluate an LLM-based Autonomous Mission Decision Support layer in which three foundation models (Llama-3.3-70B, DeepSeek-V3, and Qwen3-A22B) are queried live via the Nebius AI Studio API across ten structured anomaly scenarios derived directly from the preceding analyses. The best-performing model achieves 80% decision accuracy against physics-grounded ground truth, with all 180 inference calls completing within a 2 s latency budget consistent with radiation-hardened edge deployment, establishing the viability of foundation models as an onboard cognitive layer for high-ANS missions. △ Less

Submitted 30 March, 2026; originally announced March 2026.

AeroGrab: A Unified Framework for Aerial Grasping in Cluttered Environments

Authors: Shivansh Pratap Singh, Naveen Sudheer Nair, Samaksh Ujjawal, Sarthak Mishra, Soham Patil, Rishabh Dev Yadav, Spandan Roy

Abstract: Reliable aerial grasping in cluttered environments remains challenging due to occlusions and collision risks. Existing aerial manipulation pipelines largely rely on centroid-based grasping and lack integration between the grasp pose generation models, active exploration, and language-level task specification, resulting in the absence of a complete end-to-end system. In this work, we present an int… ▽ More Reliable aerial grasping in cluttered environments remains challenging due to occlusions and collision risks. Existing aerial manipulation pipelines largely rely on centroid-based grasping and lack integration between the grasp pose generation models, active exploration, and language-level task specification, resulting in the absence of a complete end-to-end system. In this work, we present an integrated pipeline for reliable aerial grasping in cluttered environments. Given a scene and a language instruction, the system identifies the target object and actively explores it to gain better views of the object. During exploration, a grasp generation network predicts multiple 6-DoF grasp candidates for each view. Each candidate is evaluated using a collision-aware feasibility framework, and the overall best grasp is selected and executed using standard trajectory generation and control methods. Experiments in cluttered real-world scenarios demonstrate robust and reliable grasp execution, highlighting the effectiveness of combining active perception with feasibility-aware grasp selection for aerial manipulation. △ Less

Submitted 26 June, 2026; v1 submitted 16 March, 2026; originally announced March 2026.

Nitrogen-Vacancy-Mediated Magnetism in Sputtered GdN Thin Films

Authors: Pankaj Bhardwaj, Jyotirmoy Sarkar, Bubun Biswal, Subhransu Kumar Negi, Arijit Sinha, Anirudh Venugopalrao, Sharath Kumar C, Sreelakshmi M Nair, R. S. Patel, Deepshika Jaiswal Nagar, Abhishek Mishra, Srinivasan Raghavan, Umesh Waghmare, Dhavala Suri

Abstract: Among rare-earth nitrides (RENs), gadolinium nitride (GdN) stands out as a promising material for spintronics owing to its distinctive combination of semiconducting behavior, strong exchange interactions, and intrinsically soft ferromagnetism. Its relatively high Curie temperature and large saturation magnetization make it an attractive candidate for device concepts such as non-volatile memory ele… ▽ More Among rare-earth nitrides (RENs), gadolinium nitride (GdN) stands out as a promising material for spintronics owing to its distinctive combination of semiconducting behavior, strong exchange interactions, and intrinsically soft ferromagnetism. Its relatively high Curie temperature and large saturation magnetization make it an attractive candidate for device concepts such as non-volatile memory elements and spin-based transistors, motivating efforts toward low-cost, uniform, and compositionally controlled thin-film growth. In this work, we deposited GdN thin films on SiO2/AlN substrates using DC sputtering under reactive nitridation conditions, with thicknesses varying from 18 to 180 nm, and systematically investigated their structural and magnetic properties. The films exhibit soft ferromagnetic ordering, characterized by a coercive field of approximately 200 Oe and a Curie temperature (Tc) near 70 K. Structural analysis reveals lattice distortions and local strain associated with nitrogen-vacancy defects, whose concentration varies with film thickness. Our theoretical studies establish a direct correlation between the observed Raman modes of the GdN lattice and the reduced magnetization induced by nitrogen vacancies. These vacancies give rise to defect-mediated ferromagnetism, leading to a measurable enhancement of Tc from 68 K to 82 K across the studied thickness range. The observed magnetic behavior is well described by the bound magnetic polaron (BMP) model, confirming that nitrogen vacancies are key contributors to ferromagnetic ordering while preserving the soft-magnetic character intrinsic to GdN. This study underscores the pivotal role of defect engineering in optimizing GdN thin films for spintronics applications. △ Less

Submitted 14 March, 2026; originally announced March 2026.

Report number: Volume 59, Number 31

Journal ref: Journal of Physics D: Applied Physics,2026

EnterpriseOps-Gym: Environments and Evaluations for Stateful Agentic Planning and Tool Use in Enterprise Settings

Authors: Shiva Krishna Reddy Malay, Shravan Nayak, Jishnu Sethumadhavan Nair, Sagar Davasam, Aman Tiwari, Sathwik Tejaswi Madhusudhan, Sridhar Krishna Nemala, Srinivas Sunkara, Sai Rajeswar

Abstract: Large language models are shifting from passive information providers to active agents intended for complex workflows. However, their deployment as reliable AI workers in enterprise is stalled by benchmarks that fail to capture the intricacies of professional environments, specifically, the need for long-horizon planning amidst persistent state changes and strict access protocols. In this work, we… ▽ More Large language models are shifting from passive information providers to active agents intended for complex workflows. However, their deployment as reliable AI workers in enterprise is stalled by benchmarks that fail to capture the intricacies of professional environments, specifically, the need for long-horizon planning amidst persistent state changes and strict access protocols. In this work, we introduce EnterpriseOps-Gym, a benchmark designed to evaluate agentic planning in realistic enterprise settings. Specifically, EnterpriseOps-Gym features a containerized sandbox with 164 database tables and 512 functional tools to mimic real-world search friction. Within this environment, agents are evaluated on 1,150 expert-curated tasks across eight mission-critical verticals (including Customer Service, HR, and IT). Our evaluation of 14 frontier models reveals critical limitations in state-of-the-art models: the top-performing Claude Opus 4.5 achieves only 37.4% success. Further analysis shows that providing oracle human plans improves performance by 14-35 percentage points, pinpointing strategic reasoning as the primary bottleneck. Additionally, agents frequently fail to refuse infeasible tasks (best model achieves 53.9%), leading to unintended and potentially harmful side effects. Our findings underscore that current agents are not yet ready for autonomous enterprise deployment. More broadly, EnterpriseOps-Gym provides a concrete testbed to advance the robustness of agentic planning in professional workflows. △ Less

Submitted 13 March, 2026; originally announced March 2026.

Learn Structure, Adapt on the Fly: Multi-Scale Residual Learning and Online Adaptation for Aerial Manipulators

Authors: Samaksh Ujjawal, Naveen Sudheer Nair, Shivansh Pratap Singh, Rishabh Dev Yadav, Wei Pan, Spandan Roy

Abstract: Autonomous Aerial Manipulators (AAMs) are inherently coupled, nonlinear systems that exhibit nonstationary and multiscale residual dynamics, particularly during manipulator reconfiguration and abrupt payload variations. Conventional analytical dynamic models rely on fixed parametric structures, while static data-driven model assume stationary dynamics and degrade under configuration changes and pa… ▽ More Autonomous Aerial Manipulators (AAMs) are inherently coupled, nonlinear systems that exhibit nonstationary and multiscale residual dynamics, particularly during manipulator reconfiguration and abrupt payload variations. Conventional analytical dynamic models rely on fixed parametric structures, while static data-driven model assume stationary dynamics and degrade under configuration changes and payload variations. Moreover, existing learning architectures do not explicitly factorize cross-variable coupling and multi-scale temporal effects, conflating instantaneous inertial dynamics with long-horizon regime evolution. We propose a predictive-adaptive framework for real-time residual modeling and compensation in AAMs. The core of this framework is the Factorized Dynamics Transformer (FDT), which treats physical variables as independent tokens. This design enables explicit cross-variable attention while structurally separating short-horizon inertial dependencies from long-horizon aerodynamic effects. To address deployment-time distribution shifts, a Latent Residual Adapter (LRA) performs rapid linear adaptation in the latent space via Recursive Least Squares, preserving the offline nonlinear representation without prohibitive computational overhead. The adapted residual forecast is directly integrated into a residual-compensated adaptive controller. Real-world experiments on an aerial manipulator subjected to unseen payloads demonstrate higher prediction fidelity, accelerated disturbance attenuation, and superior closed-loop tracking precision compared to state-of-the-art learning baselines, all while maintaining strict real-time feasibility. △ Less

Submitted 26 June, 2026; v1 submitted 12 March, 2026; originally announced March 2026.

Integral analysis based diagnostics of turbulence model errors in skin friction

Authors: Shyam S. Nair, Vishal A. Wadhai, Robert F. Kunz, Xiang I. A. Yang

Abstract: Error diagnostics for turbulence models have traditionally focused on engineering quantities of interest, such as the skin-friction coefficient, $C_f$, most often by comparing the predicted $C_f$ against reference data. In wall-bounded turbulent boundary layers, however, $C_f$ results from several physical mechanisms -- viscous effects, turbulence, pressure gradients, and mean-flow development --… ▽ More Error diagnostics for turbulence models have traditionally focused on engineering quantities of interest, such as the skin-friction coefficient, $C_f$, most often by comparing the predicted $C_f$ against reference data. In wall-bounded turbulent boundary layers, however, $C_f$ results from several physical mechanisms -- viscous effects, turbulence, pressure gradients, and mean-flow development -- whose relative importance depends on the flow conditions. Modeling errors in these mechanisms vary across turbulence closures, and identifying them offers valuable physical insight for model evaluation and improvement. We propose a diagnostics framework that systematically isolates and quantifies such errors using the angular momentum integral (AMI) formulation. The method is applied to five transport-type Reynolds-averaged Navier-Stokes (RANS) models in two test cases: a canonical zero-pressure-gradient flat-plate boundary layer and flow over a three-dimensional hill. For the flat-plate case, comparison with direct numerical simulation (DNS) data shows that all models reproduce $C_f$ reasonably well, but often through strong error cancellation, particularly between the turbulent torque and mean-flux contributions; individual terms can deviate by more than 20% of $C_f$. For the hill case, where wall-resolved large-eddy simulation (WRLES) is used as the reference, errors are significantly larger. The dominant erroneous contribution differs by model and may exceed several times the local $C_f$, depending on streamwise position. In separated-flow regions, the error cancellation that was observed in the flat-plate case largely disappears for the hill case, and the leading source of error shifts between mechanisms. These results highlight the value of mechanism-resolved diagnostics and provide guidance for targeted turbulence-model improvements. △ Less

Submitted 11 March, 2026; originally announced March 2026.

Comments: 28 pages, 18 figures

Understanding the formation and eruption of sigmoidal structure through data-driven modeling of magnetic evolution in solar active region 13500

Authors: P. Vemareddy, S. Nair, S. Gosain

Abstract: We investigate the magnetic origin of the coronal mass ejection (CME) that occurred on November 28, 2023, at 19:50UT from active region (AR) 13500 located near the solar disk-center. The eruption was associated with an S-shaped sigmoidal structure formed by the inner AR polarities along a sheared polarity inversion line, while the outer polarities evolved through proper motions. During November 26… ▽ More We investigate the magnetic origin of the coronal mass ejection (CME) that occurred on November 28, 2023, at 19:50UT from active region (AR) 13500 located near the solar disk-center. The eruption was associated with an S-shaped sigmoidal structure formed by the inner AR polarities along a sheared polarity inversion line, while the outer polarities evolved through proper motions. During November 26-28, the AR exhibited a decrease in net magnetic flux while progressively injecting magnetic helicity and energy into the corona toward the eruption onset, highlighting the key role of helicity-injection in triggering eruptions. To simulate this magnetic evolution, we employed a data-driven magnetofrictional (MF) simulation starting 2.8 days prior to the eruption. The energy input for the model was constrained using the observed energy injection through an ad-hoc parameter. The initial potential-field configuration gradually evolved into a sheared-arcade and eventually developed into a twisted flux rope (FR) over the observed time-scale. Proxy emission maps based on electric currents show remarkable morphological agreement between the simulated and observed sigmoidal structure. The average FR-core twist increasingly builds-up leading the FR to initiate slow-rise motion of FR top from 50Mm until its eruption onset at 80Mm. Importantly, the ratio of current-carrying to total relative-helicity increased from 0.13 at FR formation to 0.30 at eruption, when the FR core entered the torus-unstable regime, suggesting an association between torus-instability and a threshold helicity ratio. These results demonstrate that data-driven MF simulations can successfully reproduce the evolving coronal magnetic configuration and may provide a robust tool for assessing the eruptive potential of ARs, particularly the helicity ratio. △ Less

Submitted 7 March, 2026; originally announced March 2026.

Comments: Accepted to Publish in ApJ, 15 pages, 11 Figures

A framework to reason about consistency and atomicity guarantees in a sparsely-connected, partially-replicated peer-to-peer system

Authors: Sreeja S. Nair, Nicholas E. Marino, Nick Pascucci, Russell Brown, Arthur P. R. Silva, Tim Cummings, Connor M. Power

Abstract: For an offline-first collaborative application to operate in true peer-to-peer fashion, its collaborative features must function even in environments where internet connectivity is limited or unavailable. Each peer may only be interested in a subset of the application data relevant to its workload, and this subset can overlap in different ways with those of other peers. Limitations imposed by acce… ▽ More For an offline-first collaborative application to operate in true peer-to-peer fashion, its collaborative features must function even in environments where internet connectivity is limited or unavailable. Each peer may only be interested in a subset of the application data relevant to its workload, and this subset can overlap in different ways with those of other peers. Limitations imposed by access control and mesh network technologies often result in peers being sparsely connected. Reasoning about consistency in these systems is hard, especially when considering transactional updates that may alter different sets of data in the same transaction. We present \textsc{IntersectionAtomicity} and \textsc{IntersectionCC} as models to reason about offline-first collaborative applications that are sparsely-connected and rely on partially replicating different subsets of a broader set of data. We then use these models to propose a set of guidelines to help developers design their application with atomicity and consistency guarantees. △ Less

Submitted 4 March, 2026; originally announced March 2026.

Comments: 7 pages, 1 figure

MSC Class: H.3.4

MEM: Multi-Scale Embodied Memory for Vision Language Action Models

Authors: Marcel Torne, Karl Pertsch, Homer Walke, Kyle Vedder, Suraj Nair, Brian Ichter, Allen Z. Ren, Haohuan Wang, Jiaming Tang, Kyle Stachowicz, Karan Dhabalia, Michael Equi, Quan Vuong, Jost Tobias Springenberg, Sergey Levine, Chelsea Finn, Danny Driess

Abstract: Conventionally, memory in end-to-end robotic learning involves inputting a sequence of past observations into the learned policy. However, in complex multi-stage real-world tasks, the robot's memory must represent past events at multiple levels of granularity: from long-term memory that captures abstracted semantic concepts (e.g., a robot cooking dinner should remember which stages of the recipe a… ▽ More Conventionally, memory in end-to-end robotic learning involves inputting a sequence of past observations into the learned policy. However, in complex multi-stage real-world tasks, the robot's memory must represent past events at multiple levels of granularity: from long-term memory that captures abstracted semantic concepts (e.g., a robot cooking dinner should remember which stages of the recipe are already done) to short-term memory that captures recent events and compensates for occlusions (e.g., a robot remembering the object it wants to pick up once its arm occludes it). In this work, our main insight is that an effective memory architecture for long-horizon robotic control should combine multiple modalities to capture these different levels of abstraction. We introduce Multi-Scale Embodied Memory (MEM), an approach for mixed-modal long-horizon memory in robot policies. MEM combines video-based short-horizon memory, compressed via a video encoder, with text-based long-horizon memory. Together, they enable robot policies to perform tasks that span up to fifteen minutes, like cleaning up a kitchen, or preparing a grilled cheese sandwich. Additionally, we find that memory enables MEM policies to intelligently adapt manipulation strategies in-context. △ Less

Submitted 7 March, 2026; v1 submitted 3 March, 2026; originally announced March 2026.

Comments: Website: https://pi.website/research/memory

BladeSDF : Unconditional and Conditional Generative Modeling of Representative Blade Geometries Using Signed Distance Functions

Authors: Ashish S. Nair, Sandipp Krishnan Ravi, Itzel Salgado, Changjie Sun, Sayan Ghosh, Liping Wang

Abstract: Generative AI has emerged as a transformative paradigm in engineering design, enabling automated synthesis and reconstruction of complex 3D geometries while preserving feasibility and performance relevance. This paper introduces a domain-specific implicit generative framework for turbine blade geometry using DeepSDF, addressing critical gaps in performance-aware modeling and manufacturable design… ▽ More Generative AI has emerged as a transformative paradigm in engineering design, enabling automated synthesis and reconstruction of complex 3D geometries while preserving feasibility and performance relevance. This paper introduces a domain-specific implicit generative framework for turbine blade geometry using DeepSDF, addressing critical gaps in performance-aware modeling and manufacturable design generation. The proposed method leverages a continuous signed distance function (SDF) representation to reconstruct and generate smooth, watertight geometries with quantified accuracy. It establishes an interpretable, near-Gaussian latent space that aligns with blade-relevant parameters, such as taper and chord ratios, enabling controlled exploration and unconditional synthesis through interpolation and Gaussian sampling. In addition, a compact neural network maps engineering descriptors, such as maximum directional strains, to latent codes, facilitating the generation of performance-informed geometry. The framework achieves high reconstruction fidelity, with surface distance errors concentrated within $1\%$ of the maximum blade dimension, and demonstrates robust generalization to unseen designs. By integrating constraints, objectives, and performance metrics, this approach advances beyond traditional 2D-guided or unconstrained 3D pipelines, offering a practical and interpretable solution for data-driven turbine blade modeling and concept generation. △ Less

Submitted 19 January, 2026; originally announced January 2026.

Clozing the Gap: Exploring Why Language Model Surprisal Outperforms Cloze Surprisal

Authors: Sathvik Nair, Byung-Doh Oh

Abstract: How predictable a word is can be quantified in two ways: using human responses to the cloze task or using probabilities from language models (LMs).When used as predictors of processing effort, LM probabilities outperform probabilities derived from cloze data. However, it is important to establish that LM probabilities do so for the right reasons, since different predictors can lead to different sc… ▽ More How predictable a word is can be quantified in two ways: using human responses to the cloze task or using probabilities from language models (LMs).When used as predictors of processing effort, LM probabilities outperform probabilities derived from cloze data. However, it is important to establish that LM probabilities do so for the right reasons, since different predictors can lead to different scientific conclusions about the role of prediction in language comprehension. We present evidence for three hypotheses about the advantage of LM probabilities: not suffering from low resolution, distinguishing semantically similar words, and accurately assigning probabilities to low-frequency words. These results call for efforts to improve the resolution of cloze studies, coupled with experiments on whether human-like prediction is also as sensitive to the fine-grained distinctions made by LM probabilities. △ Less

Submitted 26 May, 2026; v1 submitted 14 January, 2026; originally announced January 2026.

Comments: 18 pages, 10 figures, accepted to ACL 2026 Main Conference

DP-FedSOFIM: Differentially Private Federated Stochastic Optimization using Regularized Fisher Information Matrix

Authors: Sidhant Nair, Tanmay Sen, Mrinmay Sen, Sayantan Banerjee

Abstract: Differentially private federated learning (DP-FL) often suffers from slow convergence under tight privacy budgets because the noise required for privacy preservation degrades gradient quality. Although second-order optimization can accelerate training, existing approaches for DP-FL face significant scalability limitations: Newton-type methods require clients to compute Hessians, while feature cova… ▽ More Differentially private federated learning (DP-FL) often suffers from slow convergence under tight privacy budgets because the noise required for privacy preservation degrades gradient quality. Although second-order optimization can accelerate training, existing approaches for DP-FL face significant scalability limitations: Newton-type methods require clients to compute Hessians, while feature covariance methods scale poorly with model dimension. We propose DP-FedSOFIM, a simple and scalable Hessian approximation-based second-order optimization method for DP-FL. The method constructs a regularized proxy for the Fisher information matrix at the server using only privatized aggregated gradients, capturing useful curvature information without requiring full Hessian computations or feature covariance estimation. Efficient rank-one updates based on the Sherman-Morrison formula enable communication costs proportional to the model size and require only O(d) client-side memory. Because all curvature and preconditioning operations are performed at the server on already privatized gradients, DP-FedSOFIM introduces no additional privacy cost beyond the underlying privatized gradient release mechanism. Experiments on CIFAR-10 and PathMNIST demonstrate that DP-FedSOFIM converges faster and consistently achieves higher accuracy than several competitive differentially private federated learning baselines across a wide range of privacy budgets. △ Less

Submitted 20 June, 2026; v1 submitted 14 January, 2026; originally announced January 2026.

Comments: 59 pages, 8 figures, 16 tables. Accepted to TMLR

MSC Class: 68T05; 90C15; 94A27; 68Q25 ACM Class: C.2.4; I.2.6; I.2.11; K.4.1; G.1.6

Optimized Broadband Cryogenic Ferromagnetic Resonance Spectrometer using a Closed Cycle Refrigerator

Authors: Anna Merin Francis, Sunil Nair

Abstract: We present a vector network analyzer (VNA) based broadband cryogenic ferromagnetic resonance (FMR) spectrometer, operating up to 20 GHz over a temperature range of 11 to 350 K. A cost effective architecture is implemented through the integration of a closed cycle refrigerator (CCR) and a custom fabricated grounded coplanar waveguide (GCPW), designed for broadband transmission and reliable cryogeni… ▽ More We present a vector network analyzer (VNA) based broadband cryogenic ferromagnetic resonance (FMR) spectrometer, operating up to 20 GHz over a temperature range of 11 to 350 K. A cost effective architecture is implemented through the integration of a closed cycle refrigerator (CCR) and a custom fabricated grounded coplanar waveguide (GCPW), designed for broadband transmission and reliable cryogenic operation. The VNA calibration is performed prior to measurements to account for microwave background and transmission losses, enabling reliable extraction of FMR spectra across the full temperature / frequency range. The sensitivity of the spectrometer is benchmarked using a yttrium iron garnet (YIG) thin film, yielding well resolved resonances with narrow linewidths and high sensitivity. △ Less

Submitted 10 January, 2026; originally announced January 2026.

Comments: 7 pages, 7 figures

A Critical Examination of Active Learning Workflows in Materials Science

Authors: Akhil S. Nair, Lucas Foppa

Abstract: Active learning (AL) plays a critical role in materials science, enabling applications such as the construction of machine-learning interatomic potentials for atomistic simulations and the operation of self-driving laboratories. Despite its widespread use, the reliability and effectiveness of AL workflows depend on implicit design assumptions that are rarely examined systematically. Here, we criti… ▽ More Active learning (AL) plays a critical role in materials science, enabling applications such as the construction of machine-learning interatomic potentials for atomistic simulations and the operation of self-driving laboratories. Despite its widespread use, the reliability and effectiveness of AL workflows depend on implicit design assumptions that are rarely examined systematically. Here, we critically assess AL workflows deployed in materials science and investigate how key design choices, such as surrogate models, sampling strategies, uncertainty quantification and evaluation metrics, relate to their performance. By identifying common pitfalls and discussing practical mitigation strategies, we provide guidance to practitioners for the efficient design, assessment, and interpretation of AL workflows in materials science. △ Less

Submitted 9 January, 2026; originally announced January 2026.

Succeeding at Scale: Enterprise Retrieval Benchmark Construction and Index-Preserving Query Adaptation for Multi-Tenant Search

Authors: Prateek Jain, Shabari S Nair, Ritesh Goru, Prakhar Agarwal, Ajay Yadav, Yoga Sri Varshan Varadharajan, Constantine Caramanis

Abstract: Large-scale multi-tenant retrieval systems generate extensive query logs but lack curated relevance labels for effective domain adaptation, resulting in substantial underutilized "dark data." This challenge is compounded by the high cost of model updates, as jointly fine-tuning query and document encoders requires full corpus re-indexing, which is impractical in multi-tenant settings with thousand… ▽ More Large-scale multi-tenant retrieval systems generate extensive query logs but lack curated relevance labels for effective domain adaptation, resulting in substantial underutilized "dark data." This challenge is compounded by the high cost of model updates, as jointly fine-tuning query and document encoders requires full corpus re-indexing, which is impractical in multi-tenant settings with thousands of isolated indices. We introduce DevRev-Search, a passage retrieval benchmark for technical customer support built via a fully automated pipeline. Candidate generation uses fusion across diverse sparse and dense retrievers, followed by an LLM-as-a-Judge for consistency filtering and relevance labeling. We further study and systematically evaluate index-preserving query-only adaptation strategies that fine-tune only the query-encoder while keeping the document indices fixed. Experiments on DevRev-Search, SciFact, and FiQA-2018 show that parameter-efficient fine-tuning of the query encoder delivers a remarkable quality-efficiency trade-off, enabling scalable and practical enterprise multi-tenant retrieval. △ Less

Submitted 11 June, 2026; v1 submitted 8 January, 2026; originally announced January 2026.

Chronicals: A High-Performance Framework for LLM Fine-Tuning with 3.51x Speedup over Unsloth

Authors: Arjun S. Nair

Abstract: Large language model fine-tuning is bottlenecked by memory: a 7B parameter model requires 84GB--14GB for weights, 14GB for gradients, and 56GB for FP32 optimizer states--exceeding even A100-40GB capacity. We present Chronicals, an open-source training framework achieving 3.51x speedup over Unsloth through four synergistic optimizations: (1) fused Triton kernels eliminating 75% of memory traffic vi… ▽ More Large language model fine-tuning is bottlenecked by memory: a 7B parameter model requires 84GB--14GB for weights, 14GB for gradients, and 56GB for FP32 optimizer states--exceeding even A100-40GB capacity. We present Chronicals, an open-source training framework achieving 3.51x speedup over Unsloth through four synergistic optimizations: (1) fused Triton kernels eliminating 75% of memory traffic via RMSNorm (7x), SwiGLU (5x), and QK-RoPE (2.3x) fusion; (2) Cut Cross-Entropy reducing logit memory from 5GB to 135MB through online softmax computation; (3) LoRA+ with theoretically-derived 16x differential learning rates between adapter matrices; and (4) Best-Fit Decreasing sequence packing recovering 60-75% of compute wasted on padding. On Qwen2.5-0.5B with A100-40GB, Chronicals achieves 41,184 tokens/second for full fine-tuning versus Unsloth's 11,736 tokens/second (3.51x). For LoRA at rank 32, we reach 11,699 tokens/second versus Unsloth MAX's 2,857 tokens/second (4.10x). Critically, we discovered that Unsloth's reported 46,000 tokens/second benchmark exhibited zero gradient norms--the model was not training. We provide complete mathematical foundations: online softmax correctness proofs, FlashAttention IO complexity bounds O(N^2 d^2 M^{-1}), LoRA+ learning rate derivations from gradient magnitude analysis, and bin-packing approximation guarantees. All implementations, benchmarks, and proofs are available at https://github.com/Ajwebdevs/Chronicals with pip installation via https://pypi.org/project/chronicals/. △ Less

Submitted 5 January, 2026; originally announced January 2026.

Comments: 61 pages, 25 figures, open-source framework available at https://github.com/Ajwebdevs/Chronicals and pip install chronicals

MSC Class: 68T05 ACM Class: I.2.6; I.2.7

Emergence of Human to Robot Transfer in Vision-Language-Action Models

Authors: Simar Kareer, Karl Pertsch, James Darpinian, Judy Hoffman, Danfei Xu, Sergey Levine, Chelsea Finn, Suraj Nair

Abstract: Vision-language-action (VLA) models can enable broad open world generalization, but require large and diverse datasets. It is appealing to consider whether some of this data can come from human videos, which cover diverse real-world situations and are easy to obtain. However, it is difficult to train VLAs with human videos alone, and establishing a mapping between humans and robots requires manual… ▽ More Vision-language-action (VLA) models can enable broad open world generalization, but require large and diverse datasets. It is appealing to consider whether some of this data can come from human videos, which cover diverse real-world situations and are easy to obtain. However, it is difficult to train VLAs with human videos alone, and establishing a mapping between humans and robots requires manual engineering and presents a major research challenge. Drawing inspiration from advances in large language models, where the ability to learn from diverse supervision emerges with scale, we ask whether a similar phenomenon holds for VLAs that incorporate human video data. We introduce a simple co-training recipe, and find that human-to-robot transfer emerges once the VLA is pre-trained on sufficient scenes, tasks, and embodiments. Our analysis suggests that this emergent capability arises because diverse pretraining produces embodiment-agnostic representations for human and robot data. We validate these findings through a series of experiments probing human to robot skill transfer and find that with sufficiently diverse robot pre-training our method can nearly double the performance on generalization settings seen only in human data. △ Less

Submitted 26 December, 2025; originally announced December 2025.

Emergent topological properties in spatially modulated sub-wavelength barrier lattices

Authors: Giedrius Žlabys, Wen-Bin He, Domantas Burba, Sarika Sasidharan Nair, Thomas Busch, Tomoki Ozawa

Abstract: We investigate topological phenomena in a spatially modulated Dirac-$δ$ lattice, where the scattering potential varies periodically in space. Changing the potential modulation frequency leads to Hofstadter's butterfly-like energy spectrum and enables the emergence of topological transport regimes characterized by non-trivial Chern numbers. We show how the considered modulated system is connected t… ▽ More We investigate topological phenomena in a spatially modulated Dirac-$δ$ lattice, where the scattering potential varies periodically in space. Changing the potential modulation frequency leads to Hofstadter's butterfly-like energy spectrum and enables the emergence of topological transport regimes characterized by non-trivial Chern numbers. We show how the considered modulated system is connected to the Hofstadter model via the Harper equation. By adiabatically varying spatial modulation parameters, we demonstrate controllable quantum transport and verify the topological nature of these effects through Wannier center displacement and bulk invariant calculations. We also propose an experimentally feasible realization of such a system using optically controlled three-level atoms. Our findings showcase spatially engineered Kronig-Penney-type systems as versatile platforms for investigating and exploiting different topological quantum transport regimes. △ Less

Submitted 18 December, 2025; originally announced December 2025.

AERMANI-Diffusion: Regime-Conditioned Diffusion for Dynamics Learning in Aerial Manipulators

Authors: Samaksh Ujjawal, Shivansh Pratap Singh, Naveen Sudheer Nair, Rishabh Dev Yadav, Wei Pan, Spandan Roy

Abstract: Aerial manipulators undergo rapid, configuration-dependent changes in inertial coupling forces and aerodynamic forces, making accurate dynamics modeling a core challenge for reliable control. Analytical models lose fidelity under these nonlinear and nonstationary effects, while standard data-driven methods such as deep neural networks and Gaussian processes cannot represent the diverse residual be… ▽ More Aerial manipulators undergo rapid, configuration-dependent changes in inertial coupling forces and aerodynamic forces, making accurate dynamics modeling a core challenge for reliable control. Analytical models lose fidelity under these nonlinear and nonstationary effects, while standard data-driven methods such as deep neural networks and Gaussian processes cannot represent the diverse residual behaviors that arise across different operating conditions. We propose a regime-conditioned diffusion framework that models the full distribution of residual forces using a conditional diffusion process and a lightweight temporal encoder. The encoder extracts a compact summary of recent motion and configuration, enabling consistent residual predictions even through abrupt transitions or unseen payloads. When combined with an adaptive controller, the framework enables dynamics uncertainty compensation and yields markedly improved tracking accuracy in real-world tests. △ Less

Submitted 11 December, 2025; originally announced December 2025.

Neutrino mass constraints in the context of 4-parameter dark energy equation of state and DESI DR2 observations

Authors: Gowri S Nair, Amlan Chakraborty, Luca Amendola, Subinoy Das

Abstract: Cosmological constraints on the total neutrino mass, $\sum m_ν$, are strongly shaped by assumptions about the dark-energy equation of state due to the well-known degeneracy between massive neutrinos and late-time cosmic acceleration. In this work, we move beyond the two-parameter Chevallier-Polarski-Linder (CPL) form adopted in recent DESI analyses and re-examine neutrino mass constraints using a… ▽ More Cosmological constraints on the total neutrino mass, $\sum m_ν$, are strongly shaped by assumptions about the dark-energy equation of state due to the well-known degeneracy between massive neutrinos and late-time cosmic acceleration. In this work, we move beyond the two-parameter Chevallier-Polarski-Linder (CPL) form adopted in recent DESI analyses and re-examine neutrino mass constraints using a flexible four-parameter dark energy equation of state (4pDE). We implement the 4pDE model in a modified version of CLASS and perform a full MCMC analysis using Planck, DESI DR2 BAO, and Pantheon+ data. Relative to our previous 4pDE study based on pre-DESI BAO datasets, the inclusion of DESI DR2 substantially tightens the constraints on the transition parameters while still yielding a relaxed neutrino-mass bound compared to $Λ$CDM, $\sum m_ν< 0.101$ eV ($95\%$ C.L.). This upper limit is more stringent than the DESI DR2 constraint obtained within the $w_0w_a$CDM framework. From the best-fit parameters, we reconstruct the evolution of the 4pDE equation of state along with both $68\%$ and $95\%$C.L. We do not find a statistically significant phantom-crossing at $z \sim 0.5$, consistent with the conclusion from the DESI collaboration; at higher redshifts, the reconstructed $w(z)$ follows the CPL evolution and deviates only at low redshift. Additionally we also find reduction in $Δχ^2_{\rm min}=-7.3$ compared to $Λ$CDM model. △ Less

Submitted 9 December, 2025; originally announced December 2025.

Comments: 7 pages, 3 figures

Observation of Raman anomaly and characterization of magnetic phases in van der Waals ferromagnet Fe$_5$GeTe$_2$

Authors: Sreelakshmi M. Nair, Aabhaas Vineet Mallik, R. S. Patel

Abstract: Two-dimensional (2D) van der Waals (vdW) ferromagnet Fe$_5$GeTe$_2$ has garnered significant interest due to its high Curie temperature (T$_C$), large saturation magnetization, and complex magnetic behavior arising, in part, from multiple inequivalent iron sites and vacancies. While several aspects of its complex magnetic and structural characteristics have been examined through careful experiment… ▽ More Two-dimensional (2D) van der Waals (vdW) ferromagnet Fe$_5$GeTe$_2$ has garnered significant interest due to its high Curie temperature (T$_C$), large saturation magnetization, and complex magnetic behavior arising, in part, from multiple inequivalent iron sites and vacancies. While several aspects of its complex magnetic and structural characteristics have been examined through careful experiments and first principles studies, much of it remains debatable. In this study, we present one of the first comprehensive temperature-dependent Raman spectrum for bulk Fe$_5$GeTe$_2$ and in the process reveal an interesting peak shift anomaly at 150 K. We discuss the possible relationship of this Raman anomaly with the anomalous lattice expansion reported earlier for this material at around 110 K. The impact of the anomalous lattice expansion on the magnetic anisotropy in this van der Waals material is also revealed by an isothermal magnetization analysis. These findings will prove crucial for the use of Fe$_5$GeTe$_2$ in high-performance spintronic devices. △ Less

Submitted 9 December, 2025; originally announced December 2025.