Showing 1–50 of 84 results for author: Robinson, J

AI's Capability in Assisting Scientific Research in Physics, Astrophysics, and Cosmology II: Project Planning and Proposal Evaluation

Authors: Jia Liu, Veena Krishnaraj, Kateryna Vovk, Kosuke Aizawa, Adrian E. Bayer, Linda Blot, Jessica Cowell, Suyog Garg, Jonathan Grée, Anamaria Hell, Ben Horowitz, Masaya Ichikawa, Kanyuni Iemoto, Keigo Kondo, Zacharie Lorsin, Kevin McCarthy, Jamie Robinson, Miguel Ruiz-Granda, Leander Thiele, Ievgen Vovk, Mingshen Zhou

Abstract: We investigate how well large language models (LLMs) can assist scientific project planning and proposal evaluation. One-page project plans were independently generated for eight expert-conceived research projects in physics, astrophysics, and cosmology by human researchers and three contemporary LLMs (ChatGPT, Claude, and DeepSeek; mid-2025 models, used with their default tool access). The result… ▽ More We investigate how well large language models (LLMs) can assist scientific project planning and proposal evaluation. One-page project plans were independently generated for eight expert-conceived research projects in physics, astrophysics, and cosmology by human researchers and three contemporary LLMs (ChatGPT, Claude, and DeepSeek; mid-2025 models, used with their default tool access). The resulting 32 proposals were blindly evaluated by four human reviewers and two newer frontier LLMs (Claude Opus 4.8 and ChatGPT Pro 5.5) using a four-aspect evaluation rubric. Reviewers were also asked to identify whether each proposal was written by a human or an AI. Human reviewers rated human- and AI-written proposals similarly overall, whereas both AI reviewers scored AI-written proposals about one point higher (on a five-point scale) than human-written proposals. Human reviewers correctly identified human- and AI-written proposals 72% and 79% of the time, respectively, while both AI reviewers correctly classified all 32 proposals (100%). These results suggest that current LLMs can produce project plans comparable to human-written ones in the eyes of human reviewers, but that AI reviewers show a systematic preference for AI-generated proposals. Our results suggest caution when deploying LLMs widely in proposal preparation and evaluation. △ Less

Submitted 28 July, 2026; originally announced July 2026.

Comments: 16 pages, 4 figures

AI's Capability in Assisting Scientific Research in Physics, Astrophysics, and Cosmology I: Literature Review

Authors: Anamaria Hell, Kateryna Vovk, Veena Krishnaraj, Jia Liu, Kosuke Aizawa, Adrian E. Bayer, Linda Blot, Jessica Cowell, Suyog Garg, Jonathan Grée, Ben Horowitz, Masaya Ichikawa, Kanyuni Iemoto, Keigo Kondo, Zacharie Lorsin, Kevin McCarthy, Jamie Robinson, Miguel Ruiz-Granda, Leander Thiele, Ievgen Vovk, Mingshen Zhou

Abstract: We investigate how well large language models (LLMs) can assist with literature reviews for scientific research. We perform a controlled study of eight expert-conceived research projects across the areas of physics, astrophysics, and cosmology. Each project has a defined background and goal, and human experts and AI prompters are asked to perform identical literature review tasks in parallel. We c… ▽ More We investigate how well large language models (LLMs) can assist with literature reviews for scientific research. We perform a controlled study of eight expert-conceived research projects across the areas of physics, astrophysics, and cosmology. Each project has a defined background and goal, and human experts and AI prompters are asked to perform identical literature review tasks in parallel. We compare the relevant literature selected by humans with that selected by mid-2025 LLMs (ChatGPT-4o, ChatGPT Deep Research, and Gemini). We find the overlap between human- and AI-selected references to be small ($<$6\%), indicating that AI models do not yet reproduce a competent expert search on their own, though they have the potential to complement literature searches by humans. We then assess the reliability and completeness of AI-generated candidate references, distinguishing two types of hallucination: fabrications (references to nonexistent papers) and metadata mismatches (real papers with one or more incorrect fields). We find that while fabricated references make up 3\% of the AI-generated references, 64\% are real papers with at least one incorrect field (title, author, year, journal, DOI, or link), indicating that the mid-2025 models require systematic verification. However, the performance is significantly improved for the 2026 model ChatGPT Pro 5.5, with a single-project test showing zero fabrication or metadata mismatches. △ Less

Submitted 28 July, 2026; originally announced July 2026.

Comments: 12 pages, 3 figures

Report number: IPMU26-0029

Transferable Self-Harm Surveillance from Emergency Department Triage Notes Using an Evidence-Augmented Machine Learning Approach

Authors: Liuliu Chen, Gowri Rajaram, Eleanor Bailey, Katrina Witt, Michelle Lamblin, Jo Robinson, Mike Conway, Vlada Rozova

Abstract: Self-harm is a major public health concern, but current surveillance relying on hospital presentations is inadequate due to the low sensitivity of diagnostic codes. Emergency Department (ED) triage notes, recorded at the initial point of contact, provide a succinct summary of presentations and an opportunity to identify self-harm. We developed a three-stage approach, augmenting traditional machine… ▽ More Self-harm is a major public health concern, but current surveillance relying on hospital presentations is inadequate due to the low sensitivity of diagnostic codes. Emergency Department (ED) triage notes, recorded at the initial point of contact, provide a succinct summary of presentations and an opportunity to identify self-harm. We developed a three-stage approach, augmenting traditional machine learning with large language model-based screening and evidence extraction to detect self-harm in ED triage notes. We assessed model transferability across three Australian hospitals. Our approach showed AUPRCs of 0.887 +/- 0.016 and 0.884 +/- 0.012 during internal and external validation. Prospectively, it achieved AUPRC of 0.881 +/- 0.008 at the development site, and 0.879 +/- 0.012 and 0.816 +/- 0.015 at two external sites without site-specific retraining. A key advantage of the approach is that it enables identification of the primary self-harm method with an accuracy of 95%, supporting more granular surveillance beyond binary classification. △ Less

Submitted 1 June, 2026; originally announced June 2026.

FigSIM: A Dataset for Fine-grained Suicide Severity and Figurative Language in Suicide Memes

Authors: Liuliu Chen, Elise R. Carrotte, Brian E. Chapman, Jo Robinson, Mike Conway

Abstract: Suicide memes are memes used to express suicide-related thoughts or comment on suicide-related issues. Suicide memes are increasingly common on social media, yet remain poorly understood and potentially harmful. There is an urgent need to better understand their characteristics and to develop appropriate content moderation strategies that limits users' exposure to potentially harmful content. Curr… ▽ More Suicide memes are memes used to express suicide-related thoughts or comment on suicide-related issues. Suicide memes are increasingly common on social media, yet remain poorly understood and potentially harmful. There is an urgent need to better understand their characteristics and to develop appropriate content moderation strategies that limits users' exposure to potentially harmful content. Currently, the absence of annotated datasets of suicide memes remains a key barrier to developing and evaluating automated moderation approaches. In this paper, we introduce FigSIM, the first dataset designed for fine-grained analysis of suicide memes. The dataset consists of 1049 memes, each annotated for (1) fine-grained suicide severity levels, (2) figurative phenomena (e.g., metaphors), and (3) suicide-related content (e.g., suicide method depiction). We benchmark 16 unimodal and multimodal models across three tasks: figurative language, suicide severity, and suicide-related content detection. Overall, FigSIM demonstrates that suicide memes pose unique challenges for both modeling and content moderation. Analysis revealed biases, such as underprediction of higher suicide severity levels, especially for figurative memes. The dataset (including splits used for analyses) is publicly available. Content Warning: This paper contains suicide-related content that may be triggering. △ Less

Submitted 1 June, 2026; originally announced June 2026.

Comments: Content warning: contains suicide-related content. Accepted to Findings of the Association for Computational Linguistics: ACL 2026

Why Do Self-Harm Prediction Models Struggle to Generalise? Lexical and Semantic Variations in Emergency Department Triage Notes

Authors: Liuliu Chen, Mike Conway, Jo Robinson, Vlada Rozova

Abstract: Self-harm presentations to emergency departments (EDs) are strongly associated with higher suicide risk. NLP models have shown robust performance in detecting self-harm from triage notes within single hospitals, yet performance often declines across institutions. To examine potential causes, we compare ED triage notes from two hospitals by analyzing lexical characteristics, highly associated predi… ▽ More Self-harm presentations to emergency departments (EDs) are strongly associated with higher suicide risk. NLP models have shown robust performance in detecting self-harm from triage notes within single hospitals, yet performance often declines across institutions. To examine potential causes, we compare ED triage notes from two hospitals by analyzing lexical characteristics, highly associated predictive features, and salient topics. Our results reveal variation in lexical expression and feature importance related to self-harm across hospitals, despite consistent core themes such as self-poisoning and self-injury. These documentation differences are associated with reduced cross-site performance. Our findings provide insight into how institutional variation affects the identification of self-harm in clinical text and highlight potential methods to improve model generalisability. △ Less

Submitted 1 June, 2026; originally announced June 2026.

Comments: Accepted to CLPsych2026

Rapidly deploying on-device eye tracking by distilling visual foundation models

Authors: Cheng Jiang, Jogendra Kundu, David Colmenares, Fengting Yang, Joseph P Robinson, Ali Behrooz, Yatong An

Abstract: Eye tracking (ET) plays a critical role in augmented and virtual reality applications. However, rapidly deploying high-accuracy, on-device gaze estimation for new products remains challenging because hardware configurations (e.g., camera placement, camera pose, and illumination) often change across device generations. Visual foundation models (VFMs) excel on natural-image benchmarks and offer a pr… ▽ More Eye tracking (ET) plays a critical role in augmented and virtual reality applications. However, rapidly deploying high-accuracy, on-device gaze estimation for new products remains challenging because hardware configurations (e.g., camera placement, camera pose, and illumination) often change across device generations. Visual foundation models (VFMs) excel on natural-image benchmarks and offer a promising path to rapid training and deployment; yet, we find that off-the-shelf VFMs still struggle to reach high accuracy on specialized near-eye infrared images. To close this gap, we introduce DistillGaze, a framework that distills a VFM using labeled synthetic data and unlabeled real data for rapid, high-accuracy on-device gaze estimation. DistillGaze proceeds in two stages. First, we adapt a VFM into a domain-specialized teacher using synthetic gaze labels and unlabeled real images. Synthetic data provide scalable, high-quality gaze supervision, while unlabeled real data bridges the synthetic-to-real domain gap. Second, we train an on-device student from both teacher guidance and self-training. Evaluated on a large-scale crowd-sourced dataset spanning more than 2,000 participants, DistillGaze reduces median gaze error by 58.6% relative to synthetic-only baselines while maintaining a lightweight 256K-parameter model suitable for real-time on-device deployment. More broadly, DistillGaze offers an efficient path to training and deploying ET models that adapt to hardware changes, and a recipe for combining synthetic supervision with unlabeled real data in on-device regression tasks. △ Less

Submitted 2 July, 2026; v1 submitted 2 April, 2026; originally announced April 2026.

Validating Political Position Predictions of Arguments

Authors: Jordan Robinson, Angus R. Williams, Katie Atkinson, Anthony G. Cohn

Abstract: Real-world knowledge representation often requires capturing subjective, continuous attributes -- such as political positions -- that conflict with pairwise validation, the widely accepted gold standard for human evaluation. We address this challenge through a dual-scale validation framework applied to political stance prediction in argumentative discourse, combining pointwise and pairwise human a… ▽ More Real-world knowledge representation often requires capturing subjective, continuous attributes -- such as political positions -- that conflict with pairwise validation, the widely accepted gold standard for human evaluation. We address this challenge through a dual-scale validation framework applied to political stance prediction in argumentative discourse, combining pointwise and pairwise human annotation. Using 22 language models, we construct a large-scale knowledge base of political position predictions for 23,228 arguments drawn from 30 debates that appeared on the UK politicial television programme \textit{Question Time}. Pointwise evaluation shows moderate human-model agreement (Krippendorff's $α=0.578$), reflecting intrinsic subjectivity, while pairwise validation reveals substantially stronger alignment between human- and model-derived rankings ($α=0.86$ for the best model). This work contributes: (i) a practical validation methodology for subjective continuous knowledge that balances scalability with reliability; (ii) a validated structured argumentation knowledge base enabling graph-based reasoning and retrieval-augmented generation in political domains; and (iii) evidence that ordinal structure can be extracted from pointwise language models predictions from inherently subjective real-world discourse, advancing knowledge representation capabilities for domains where traditional symbolic or categorical approaches are insufficient. △ Less

Submitted 20 February, 2026; originally announced February 2026.

Comments: 13 pages, 6 figures, 6 tables. Under review

FastNet: Improving the physical consistency of machine-learning weather prediction models through loss function design

Authors: Tom Dunstan, Oliver Strickson, Thusal Bennett, Jack Bowyer, Matthew Burnand, James Chappell, Alejandro Coca-Castro, Kirstine Ida Dale, Eric G. Daub, Noushin Eftekhari, Manvendra Janmaijaya, Jon Lillis, David Salvador-Jasin, Nathan Simpson, Ryan Sze-Yin Chan, Mohamad Elmasri, Lydia Allegranza France, Sam Madge, Levan Bokeria, Hannah Brown, Tom Dodds, Anna-Louise Ellis, David Llewellyn-Jones, Theo McCaie, Sophia Moreton , et al. (9 additional authors not shown)

Abstract: Machine learning weather prediction (MLWP) models have demonstrated remarkable potential in delivering accurate forecasts at significantly reduced computational cost compared to traditional numerical weather prediction (NWP) systems. However, challenges remain in ensuring the physical consistency of MLWP outputs, particularly in deterministic settings. This study presents FastNet, a graph neural n… ▽ More Machine learning weather prediction (MLWP) models have demonstrated remarkable potential in delivering accurate forecasts at significantly reduced computational cost compared to traditional numerical weather prediction (NWP) systems. However, challenges remain in ensuring the physical consistency of MLWP outputs, particularly in deterministic settings. This study presents FastNet, a graph neural network (GNN)-based global prediction model, and investigates the impact of alternative loss function designs on improving the physical realism of its forecasts. We explore three key modifications to the standard mean squared error (MSE) loss: (1) a modified spherical harmonic (MSH) loss that penalises spectral amplitude errors to reduce blurring and enhance small-scale structure retention; (2) inclusion of horizontal gradient terms in the loss to suppress non-physical artefacts; and (3) an alternative wind representation that decouples speed and direction to better capture extreme wind events. Results show that while the MSH and gradient-based losses \textit{alone} may slightly degrade RMSE scores, when trained in combination the model exhibits very similar MSE performance to an MSE-trained model while at the same time significantly improving spectral fidelity and physical consistency. The alternative wind representation further improves wind speed accuracy and reduces directional bias. Collectively, these findings highlight the importance of loss function design as a mechanism for embedding domain knowledge into MLWP models and advancing their operational readiness. △ Less

Submitted 22 September, 2025; originally announced September 2025.

Simple and Effective Baselines for Code Summarisation Evaluation

Authors: Jade Robinson, Jonathan K. Kummerfeld

Abstract: Code documentation is useful, but writing it is time-consuming. Different techniques for generating code summaries have emerged, but comparing them is difficult because human evaluation is expensive and automatic metrics are unreliable. In this paper, we introduce a simple new baseline in which we ask an LLM to give an overall score to a summary. Unlike n-gram and embedding-based baselines, our ap… ▽ More Code documentation is useful, but writing it is time-consuming. Different techniques for generating code summaries have emerged, but comparing them is difficult because human evaluation is expensive and automatic metrics are unreliable. In this paper, we introduce a simple new baseline in which we ask an LLM to give an overall score to a summary. Unlike n-gram and embedding-based baselines, our approach is able to consider the code when giving a score. This allows us to also make a variant that does not consider the reference summary at all, which could be used for other tasks, e.g., to evaluate the quality of documentation in code bases. We find that our method is as good or better than prior metrics, though we recommend using it in conjunction with embedding-based methods to avoid the risk of LLM-specific bias. △ Less

Submitted 25 May, 2025; originally announced May 2025.

MSC Class: 68T50 ACM Class: I.2.7

Humanity's Last Exam

Authors: Long Phan, Alice Gatti, Ziwen Han, Nathaniel Li, Josephina Hu, Hugh Zhang, Chen Bo Calvin Zhang, Mohamed Shaaban, John Ling, Sean Shi, Michael Choi, Anish Agrawal, Arnav Chopra, Adam Khoja, Ryan Kim, Richard Ren, Jason Hausenloy, Oliver Zhang, Mantas Mazeika, Dmitry Dodonov, Tung Nguyen, Jaeho Lee, Daron Anderson, Mikhail Doroshenko, Alun Cennyth Stokes , et al. (1133 additional authors not shown)

Abstract: Benchmarks are important tools for tracking the rapid advancements in large language model (LLM) capabilities. However, benchmarks are not keeping pace in difficulty: LLMs now achieve over 90\% accuracy on popular benchmarks like MMLU, limiting informed measurement of state-of-the-art LLM capabilities. In response, we introduce Humanity's Last Exam (HLE), a multi-modal benchmark at the frontier of… ▽ More Benchmarks are important tools for tracking the rapid advancements in large language model (LLM) capabilities. However, benchmarks are not keeping pace in difficulty: LLMs now achieve over 90\% accuracy on popular benchmarks like MMLU, limiting informed measurement of state-of-the-art LLM capabilities. In response, we introduce Humanity's Last Exam (HLE), a multi-modal benchmark at the frontier of human knowledge, designed to be the final closed-ended academic benchmark of its kind with broad subject coverage. HLE consists of 2,500 questions across dozens of subjects, including mathematics, humanities, and the natural sciences. HLE is developed globally by subject-matter experts and consists of multiple-choice and short-answer questions suitable for automated grading. Each question has a known solution that is unambiguous and easily verifiable, but cannot be quickly answered via internet retrieval. State-of-the-art LLMs demonstrate low accuracy and calibration on HLE, highlighting a significant gap between current LLM capabilities and the expert human frontier on closed-ended academic questions. To inform research and policymaking upon a clear understanding of model capabilities, we publicly release HLE at https://lastexam.ai. △ Less

Submitted 28 July, 2026; v1 submitted 24 January, 2025; originally announced January 2025.

Comments: 29 pages, 6 figures

LocateBench: Evaluating the Locating Ability of Vision Language Models

Authors: Ting-Rui Chiang, Joshua Robinson, Xinyan Velocity Yu, Dani Yogatama

Abstract: The ability to locate an object in an image according to natural language instructions is crucial for many real-world applications. In this work we propose LocateBench, a high-quality benchmark dedicated to evaluating this ability. We experiment with multiple prompting approaches, and measure the accuracy of several large vision language models. We find that even the accuracy of the strongest mode… ▽ More The ability to locate an object in an image according to natural language instructions is crucial for many real-world applications. In this work we propose LocateBench, a high-quality benchmark dedicated to evaluating this ability. We experiment with multiple prompting approaches, and measure the accuracy of several large vision language models. We find that even the accuracy of the strongest model, GPT-4o, lags behind human accuracy by more than 10%. △ Less

Submitted 17 October, 2024; originally announced October 2024.

Comments: We release the dataset at https://usc-tamagotchi.github.io/locate-bench/

RelBench: A Benchmark for Deep Learning on Relational Databases

Authors: Joshua Robinson, Rishabh Ranjan, Weihua Hu, Kexin Huang, Jiaqi Han, Alejandro Dobles, Matthias Fey, Jan E. Lenssen, Yiwen Yuan, Zecheng Zhang, Xinwei He, Jure Leskovec

Abstract: We present RelBench, a public benchmark for solving predictive tasks over relational databases with graph neural networks. RelBench provides databases and tasks spanning diverse domains and scales, and is intended to be a foundational infrastructure for future research. We use RelBench to conduct the first comprehensive study of Relational Deep Learning (RDL) (Fey et al., 2024), which combines gra… ▽ More We present RelBench, a public benchmark for solving predictive tasks over relational databases with graph neural networks. RelBench provides databases and tasks spanning diverse domains and scales, and is intended to be a foundational infrastructure for future research. We use RelBench to conduct the first comprehensive study of Relational Deep Learning (RDL) (Fey et al., 2024), which combines graph neural network predictive models with (deep) tabular models that extract initial entity-level representations from raw tables. End-to-end learned RDL models fully exploit the predictive signal encoded in primary-foreign key links, marking a significant shift away from the dominant paradigm of manual feature engineering combined with tabular models. To thoroughly evaluate RDL against this prior gold-standard, we conduct an in-depth user study where an experienced data scientist manually engineers features for each task. In this study, RDL learns better models whilst reducing human work needed by more than an order of magnitude. This demonstrates the power of deep learning for solving predictive tasks over relational databases, opening up many new research opportunities enabled by RelBench. △ Less

Submitted 29 July, 2024; originally announced July 2024.

Surface-based parcellation and vertex-wise analysis of ultra high-resolution ex vivo 7 tesla MRI in Alzheimer's disease and related dementias

Authors: Pulkit Khandelwal, Michael Tran Duong, Lisa Levorse, Constanza Fuentes, Amanda Denning, Winifred Trotman, Ranjit Ittyerah, Alejandra Bahena, Theresa Schuck, Marianna Gabrielyan, Karthik Prabhakaran, Daniel Ohm, Gabor Mizsei, John Robinson, Monica Munoz, John Detre, Edward Lee, David Irwin, Corey McMillan, M. Dylan Tisdall, Sandhitsu Das, David Wolk, Paul A. Yushkevich

Abstract: Magnetic resonance imaging (MRI) is the standard modality to understand human brain structure and function in vivo (antemortem). Decades of research in human neuroimaging has led to the widespread development of methods and tools to provide automated volume-based segmentations and surface-based parcellations which help localize brain functions to specialized anatomical regions. Recently ex vivo (p… ▽ More Magnetic resonance imaging (MRI) is the standard modality to understand human brain structure and function in vivo (antemortem). Decades of research in human neuroimaging has led to the widespread development of methods and tools to provide automated volume-based segmentations and surface-based parcellations which help localize brain functions to specialized anatomical regions. Recently ex vivo (postmortem) imaging of the brain has opened-up avenues to study brain structure at sub-millimeter ultra high-resolution revealing details not possible to observe with in vivo MRI. Unfortunately, there has been limited methodological development in ex vivo MRI primarily due to lack of datasets and limited centers with such imaging resources. Therefore, in this work, we present one-of-its-kind dataset of 82 ex vivo T2w whole brain hemispheres MRI at 0.3 mm isotropic resolution spanning Alzheimer's disease and related dementias. We adapted and developed a fast and easy-to-use automated surface-based pipeline to parcellate, for the first time, ultra high-resolution ex vivo brain tissue at the native subject space resolution using the Desikan-Killiany-Tourville (DKT) brain atlas. This allows us to perform vertex-wise analysis in the template space and thereby link morphometry measures with pathology measurements derived from histology. We will open-source our dataset docker container, Jupyter notebooks for ready-to-use out-of-the-box set of tools and command line options to advance ex vivo MRI clinical brain imaging research on the project webpage. △ Less

Submitted 2 July, 2024; v1 submitted 28 March, 2024; originally announced March 2024.

StarCoder 2 and The Stack v2: The Next Generation

Authors: Anton Lozhkov, Raymond Li, Loubna Ben Allal, Federico Cassano, Joel Lamy-Poirier, Nouamane Tazi, Ao Tang, Dmytro Pykhtar, Jiawei Liu, Yuxiang Wei, Tianyang Liu, Max Tian, Denis Kocetkov, Arthur Zucker, Younes Belkada, Zijian Wang, Qian Liu, Dmitry Abulkhanov, Indraneil Paul, Zhuang Li, Wen-Ding Li, Megan Risdal, Jia Li, Jian Zhu, Terry Yue Zhuo , et al. (41 additional authors not shown)

Abstract: The BigCode project, an open-scientific collaboration focused on the responsible development of Large Language Models for Code (Code LLMs), introduces StarCoder2. In partnership with Software Heritage (SWH), we build The Stack v2 on top of the digital commons of their source code archive. Alongside the SWH repositories spanning 619 programming languages, we carefully select other high-quality data… ▽ More The BigCode project, an open-scientific collaboration focused on the responsible development of Large Language Models for Code (Code LLMs), introduces StarCoder2. In partnership with Software Heritage (SWH), we build The Stack v2 on top of the digital commons of their source code archive. Alongside the SWH repositories spanning 619 programming languages, we carefully select other high-quality data sources, such as GitHub pull requests, Kaggle notebooks, and code documentation. This results in a training set that is 4x larger than the first StarCoder dataset. We train StarCoder2 models with 3B, 7B, and 15B parameters on 3.3 to 4.3 trillion tokens and thoroughly evaluate them on a comprehensive set of Code LLM benchmarks. We find that our small model, StarCoder2-3B, outperforms other Code LLMs of similar size on most benchmarks, and also outperforms StarCoderBase-15B. Our large model, StarCoder2- 15B, significantly outperforms other models of comparable size. In addition, it matches or outperforms CodeLlama-34B, a model more than twice its size. Although DeepSeekCoder- 33B is the best-performing model at code completion for high-resource languages, we find that StarCoder2-15B outperforms it on math and code reasoning benchmarks, as well as several low-resource languages. We make the model weights available under an OpenRAIL license and ensure full transparency regarding the training data by releasing the SoftWare Heritage persistent IDentifiers (SWHIDs) of the source code data. △ Less

Submitted 29 February, 2024; originally announced February 2024.

Relational Deep Learning: Graph Representation Learning on Relational Databases

Authors: Matthias Fey, Weihua Hu, Kexin Huang, Jan Eric Lenssen, Rishabh Ranjan, Joshua Robinson, Rex Ying, Jiaxuan You, Jure Leskovec

Abstract: Much of the world's most valued data is stored in relational databases and data warehouses, where the data is organized into many tables connected by primary-foreign key relations. However, building machine learning models using this data is both challenging and time consuming. The core problem is that no machine learning method is capable of learning on multiple tables interconnected by primary-f… ▽ More Much of the world's most valued data is stored in relational databases and data warehouses, where the data is organized into many tables connected by primary-foreign key relations. However, building machine learning models using this data is both challenging and time consuming. The core problem is that no machine learning method is capable of learning on multiple tables interconnected by primary-foreign key relations. Current methods can only learn from a single table, so the data must first be manually joined and aggregated into a single training table, the process known as feature engineering. Feature engineering is slow, error prone and leads to suboptimal models. Here we introduce an end-to-end deep representation learning approach to directly learn on data laid out across multiple tables. We name our approach Relational Deep Learning (RDL). The core idea is to view relational databases as a temporal, heterogeneous graph, with a node for each row in each table, and edges specified by primary-foreign key links. Message Passing Graph Neural Networks can then automatically learn across the graph to extract representations that leverage all input data, without any manual feature engineering. Relational Deep Learning leads to more accurate models that can be built much faster. To facilitate research in this area, we develop RelBench, a set of benchmark datasets and an implementation of Relational Deep Learning. The data covers a wide spectrum, from discussions on Stack Exchange to book reviews on the Amazon Product Catalog. Overall, we define a new research area that generalizes graph machine learning and broadens its applicability to a wide set of AI use cases. △ Less

Submitted 7 December, 2023; originally announced December 2023.

Comments: https://relbench.stanford.edu

The BigCode Project Governance Card

Authors: BigCode collaboration, Sean Hughes, Harm de Vries, Jennifer Robinson, Carlos Muñoz Ferrandis, Loubna Ben Allal, Leandro von Werra, Jennifer Ding, Sebastien Paquet, Yacine Jernite

Abstract: This document serves as an overview of the different mechanisms and areas of governance in the BigCode project. It aims to support transparency by providing relevant information about choices that were made during the project to the broader public, and to serve as an example of intentional governance of an open research project that future endeavors can leverage to shape their own approach. The fi… ▽ More This document serves as an overview of the different mechanisms and areas of governance in the BigCode project. It aims to support transparency by providing relevant information about choices that were made during the project to the broader public, and to serve as an example of intentional governance of an open research project that future endeavors can leverage to shape their own approach. The first section, Project Structure, covers the project organization, its stated goals and values, its internal decision processes, and its funding and resources. The second section, Data and Model Governance, covers decisions relating to the questions of data subject consent, privacy, and model release. △ Less

Submitted 6 December, 2023; originally announced December 2023.

Comments: 12 pages, related papers arXiv:2305.06161 and arXiv:2301.03988 and arXiv:2211.15533v1, learn more at https://www.bigcode-project.org/

Expressive Sign Equivariant Networks for Spectral Geometric Learning

Authors: Derek Lim, Joshua Robinson, Stefanie Jegelka, Haggai Maron

Abstract: Recent work has shown the utility of developing machine learning models that respect the structure and symmetries of eigenvectors. These works promote sign invariance, since for any eigenvector v the negation -v is also an eigenvector. However, we show that sign invariance is theoretically limited for tasks such as building orthogonally equivariant models and learning node positional encodings for… ▽ More Recent work has shown the utility of developing machine learning models that respect the structure and symmetries of eigenvectors. These works promote sign invariance, since for any eigenvector v the negation -v is also an eigenvector. However, we show that sign invariance is theoretically limited for tasks such as building orthogonally equivariant models and learning node positional encodings for link prediction in graphs. In this work, we demonstrate the benefits of sign equivariance for these tasks. To obtain these benefits, we develop novel sign equivariant neural network architectures. Our models are based on a new analytic characterization of sign equivariant polynomials and thus inherit provable expressiveness properties. Controlled synthetic experiments show that our networks can achieve the theoretically predicted benefits of sign equivariant models. Code is available at https://github.com/cptq/Sign-Equivariant-Nets. △ Less

Submitted 4 December, 2023; originally announced December 2023.

Comments: NeurIPS 2023 Spotlight

On Retrieval Augmentation and the Limitations of Language Model Training

Authors: Ting-Rui Chiang, Xinyan Velocity Yu, Joshua Robinson, Ollie Liu, Isabelle Lee, Dani Yogatama

Abstract: Augmenting a language model (LM) with $k$-nearest neighbors ($k$NN) retrieval on its training data alone can decrease its perplexity, though the underlying reasons for this remain elusive. In this work, we rule out one previously posited possibility -- the "softmax bottleneck." We then create a new dataset to evaluate LM generalization ability in the setting where training data contains additional… ▽ More Augmenting a language model (LM) with $k$-nearest neighbors ($k$NN) retrieval on its training data alone can decrease its perplexity, though the underlying reasons for this remain elusive. In this work, we rule out one previously posited possibility -- the "softmax bottleneck." We then create a new dataset to evaluate LM generalization ability in the setting where training data contains additional information that is not causally relevant. This task is challenging even for GPT-3.5 Turbo. We show that, for both GPT-2 and Mistral 7B, $k$NN retrieval augmentation consistently improves performance in this setting. Finally, to make $k$NN retrieval more accessible, we propose using a multi-layer perceptron model that maps datastore keys to values as a drop-in replacement for traditional retrieval. This reduces storage costs by over 25x. △ Less

Submitted 2 April, 2024; v1 submitted 16 November, 2023; originally announced November 2023.

Comments: Accepted to NAACL 2024

On the Stability of Expressive Positional Encodings for Graphs

Authors: Yinan Huang, William Lu, Joshua Robinson, Yu Yang, Muhan Zhang, Stefanie Jegelka, Pan Li

Abstract: Designing effective positional encodings for graphs is key to building powerful graph transformers and enhancing message-passing graph neural networks. Although widespread, using Laplacian eigenvectors as positional encodings faces two fundamental challenges: (1) \emph{Non-uniqueness}: there are many different eigendecompositions of the same Laplacian, and (2) \emph{Instability}: small perturbatio… ▽ More Designing effective positional encodings for graphs is key to building powerful graph transformers and enhancing message-passing graph neural networks. Although widespread, using Laplacian eigenvectors as positional encodings faces two fundamental challenges: (1) \emph{Non-uniqueness}: there are many different eigendecompositions of the same Laplacian, and (2) \emph{Instability}: small perturbations to the Laplacian could result in completely different eigenspaces, leading to unpredictable changes in positional encoding. Despite many attempts to address non-uniqueness, most methods overlook stability, leading to poor generalization on unseen graph structures. We identify the cause of instability to be a ``hard partition'' of eigenspaces. Hence, we introduce Stable and Expressive Positional Encodings (SPE), an architecture for processing eigenvectors that uses eigenvalues to ``softly partition'' eigenspaces. SPE is the first architecture that is (1) provably stable, and (2) universally expressive for basis invariant functions whilst respecting all symmetries of eigenvectors. Besides guaranteed stability, we prove that SPE is at least as expressive as existing methods, and highly capable of counting graph structures. Finally, we evaluate the effectiveness of our method on molecular property prediction, and out-of-distribution generalization tasks, finding improved generalization compared to existing positional encoding methods. Our code is available at \url{https://github.com/Graph-COM/SPE}. △ Less

Submitted 8 June, 2024; v1 submitted 4 October, 2023; originally announced October 2023.

Comments: ICLR 2024

Structuring Representation Geometry with Rotationally Equivariant Contrastive Learning

Authors: Sharut Gupta, Joshua Robinson, Derek Lim, Soledad Villar, Stefanie Jegelka

Abstract: Self-supervised learning converts raw perceptual data such as images to a compact space where simple Euclidean distances measure meaningful variations in data. In this paper, we extend this formulation by adding additional geometric structure to the embedding space by enforcing transformations of input space to correspond to simple (i.e., linear) transformations of embedding space. Specifically, i… ▽ More Self-supervised learning converts raw perceptual data such as images to a compact space where simple Euclidean distances measure meaningful variations in data. In this paper, we extend this formulation by adding additional geometric structure to the embedding space by enforcing transformations of input space to correspond to simple (i.e., linear) transformations of embedding space. Specifically, in the contrastive learning setting, we introduce an equivariance objective and theoretically prove that its minima forces augmentations on input space to correspond to rotations on the spherical embedding space. We show that merely combining our equivariant loss with a non-collapse term results in non-trivial representations, without requiring invariance to data augmentations. Optimal performance is achieved by also encouraging approximate invariance, where input augmentations correspond to small rotations. Our method, CARE: Contrastive Augmentation-induced Rotational Equivariance, leads to improved performance on downstream tasks, and ensures sensitivity in embedding space to important variations in data (e.g., color) that standard contrastive methods do not achieve. Code is available at https://github.com/Sharut/CARE. △ Less

Submitted 24 June, 2023; originally announced June 2023.

Comments: 22 pages

StarCoder: may the source be with you!

Authors: Raymond Li, Loubna Ben Allal, Yangtian Zi, Niklas Muennighoff, Denis Kocetkov, Chenghao Mou, Marc Marone, Christopher Akiki, Jia Li, Jenny Chim, Qian Liu, Evgenii Zheltonozhskii, Terry Yue Zhuo, Thomas Wang, Olivier Dehaene, Mishig Davaadorj, Joel Lamy-Poirier, João Monteiro, Oleh Shliazhko, Nicolas Gontier, Nicholas Meade, Armel Zebaze, Ming-Ho Yee, Logesh Kumar Umapathi, Jian Zhu , et al. (42 additional authors not shown)

Abstract: The BigCode community, an open-scientific collaboration working on the responsible development of Large Language Models for Code (Code LLMs), introduces StarCoder and StarCoderBase: 15.5B parameter models with 8K context length, infilling capabilities and fast large-batch inference enabled by multi-query attention. StarCoderBase is trained on 1 trillion tokens sourced from The Stack, a large colle… ▽ More The BigCode community, an open-scientific collaboration working on the responsible development of Large Language Models for Code (Code LLMs), introduces StarCoder and StarCoderBase: 15.5B parameter models with 8K context length, infilling capabilities and fast large-batch inference enabled by multi-query attention. StarCoderBase is trained on 1 trillion tokens sourced from The Stack, a large collection of permissively licensed GitHub repositories with inspection tools and an opt-out process. We fine-tuned StarCoderBase on 35B Python tokens, resulting in the creation of StarCoder. We perform the most comprehensive evaluation of Code LLMs to date and show that StarCoderBase outperforms every open Code LLM that supports multiple programming languages and matches or outperforms the OpenAI code-cushman-001 model. Furthermore, StarCoder outperforms every model that is fine-tuned on Python, can be prompted to achieve 40\% pass@1 on HumanEval, and still retains its performance on other programming languages. We take several important steps towards a safe open-access model release, including an improved PII redaction pipeline and a novel attribution tracing tool, and make the StarCoder models publicly available under a more commercially viable version of the Open Responsible AI Model license. △ Less

Submitted 13 December, 2023; v1 submitted 9 May, 2023; originally announced May 2023.

Algorithms for Reconstructing DDoS Attack Graphs using Probabilistic Packet Marking

Authors: Dina Barak-Pelleg, Daniel Berend, Thomas J. Robinson, Itamar Zimmerman

Abstract: DoS and DDoS attacks are widely used and pose a constant threat. Here we explore Probability Packet Marking (PPM), one of the important methods for reconstructing the attack-graph and detect the attackers. We present two algorithms. Differently from others, their stopping time is not fixed a priori. It rather depends on the actual distance of the attacker from the victim. Our first algorithm retur… ▽ More DoS and DDoS attacks are widely used and pose a constant threat. Here we explore Probability Packet Marking (PPM), one of the important methods for reconstructing the attack-graph and detect the attackers. We present two algorithms. Differently from others, their stopping time is not fixed a priori. It rather depends on the actual distance of the attacker from the victim. Our first algorithm returns the graph at the earliest feasible time, and turns out to guarantee high success probability. The second algorithm enables attaining any predetermined success probability at the expense of a longer runtime. We study the performance of the two algorithms theoretically, and compare them to other algorithms by simulation. Finally, we consider the order in which the marks corresponding to the various edges of the attack graph are obtained by the victim. We show that, although edges closer to the victim tend to be discovered earlier in the process than farther edges, the differences are much smaller than previously thought. △ Less

Submitted 11 April, 2023; originally announced April 2023.

Comments: 30 pages, 4 figures, 4 tables

MSC Class: 60C05 ACM Class: G.3; I.6.6

Automated deep learning segmentation of high-resolution 7 T postmortem MRI for quantitative analysis of structure-pathology correlations in neurodegenerative diseases

Authors: Pulkit Khandelwal, Michael Tran Duong, Shokufeh Sadaghiani, Sydney Lim, Amanda Denning, Eunice Chung, Sadhana Ravikumar, Sanaz Arezoumandan, Claire Peterson, Madigan Bedard, Noah Capp, Ranjit Ittyerah, Elyse Migdal, Grace Choi, Emily Kopp, Bridget Loja, Eusha Hasan, Jiacheng Li, Alejandra Bahena, Karthik Prabhakaran, Gabor Mizsei, Marianna Gabrielyan, Theresa Schuck, Winifred Trotman, John Robinson , et al. (12 additional authors not shown)

Abstract: Postmortem MRI allows brain anatomy to be examined at high resolution and to link pathology measures with morphometric measurements. However, automated segmentation methods for brain mapping in postmortem MRI are not well developed, primarily due to limited availability of labeled datasets, and heterogeneity in scanner hardware and acquisition protocols. In this work, we present a high resolution… ▽ More Postmortem MRI allows brain anatomy to be examined at high resolution and to link pathology measures with morphometric measurements. However, automated segmentation methods for brain mapping in postmortem MRI are not well developed, primarily due to limited availability of labeled datasets, and heterogeneity in scanner hardware and acquisition protocols. In this work, we present a high resolution of 135 postmortem human brain tissue specimens imaged at 0.3 mm$^{3}$ isotropic using a T2w sequence on a 7T whole-body MRI scanner. We developed a deep learning pipeline to segment the cortical mantle by benchmarking the performance of nine deep neural architectures, followed by post-hoc topological correction. We then segment four subcortical structures (caudate, putamen, globus pallidus, and thalamus), white matter hyperintensities, and the normal appearing white matter. We show generalizing capabilities across whole brain hemispheres in different specimens, and also on unseen images acquired at 0.28 mm^3 and 0.16 mm^3 isotropic T2*w FLASH sequence at 7T. We then compute localized cortical thickness and volumetric measurements across key regions, and link them with semi-quantitative neuropathological ratings. Our code, Jupyter notebooks, and the containerized executables are publicly available at: https://pulkit-khandelwal.github.io/exvivo-brain-upenn △ Less

Submitted 17 October, 2023; v1 submitted 21 March, 2023; originally announced March 2023.

Comments: Preprint submitted to NeuroImage Project website: https://pulkit-khandelwal.github.io/exvivo-brain-upenn

Deep Learning Pipeline for Preprocessing and Segmenting Cardiac Magnetic Resonance of Single Ventricle Patients from an Image Registry

Authors: Tina Yao, Nicole St. Clair, Gabriel F. Miller, Adam L. Dorfman, Mark A. Fogel, Sunil Ghelani, Rajesh Krishnamurthy, Christopher Z. Lam, Joshua D. Robinson, David Schidlow, Timothy C. Slesnick, Justin Weigand, Michael Quail, Rahul Rathod, Jennifer A. Steeden, Vivek Muthurangu

Abstract: Purpose: To develop and evaluate an end-to-end deep learning pipeline for segmentation and analysis of cardiac magnetic resonance images to provide core-lab processing for a multi-centre registry of Fontan patients. Materials and Methods: This retrospective study used training (n = 175), validation (n = 25) and testing (n = 50) cardiac magnetic resonance image exams collected from 13 institution… ▽ More Purpose: To develop and evaluate an end-to-end deep learning pipeline for segmentation and analysis of cardiac magnetic resonance images to provide core-lab processing for a multi-centre registry of Fontan patients. Materials and Methods: This retrospective study used training (n = 175), validation (n = 25) and testing (n = 50) cardiac magnetic resonance image exams collected from 13 institutions in the UK, US and Canada. The data was used to train and evaluate a pipeline containing three deep-learning models. The pipeline's performance was assessed on the Dice and IoU score between the automated and reference standard manual segmentation. Cardiac function values were calculated from both the automated and manual segmentation and evaluated using Bland-Altman analysis and paired t-tests. The overall pipeline was further evaluated qualitatively on 475 unseen patient exams. Results: For the 50 testing dataset, the pipeline achieved a median Dice score of 0.91 (0.89-0.94) for end-diastolic volume, 0.86 (0.82-0.89) for end-systolic volume, and 0.74 (0.70-0.77) for myocardial mass. The deep learning-derived end-diastolic volume, end-systolic volume, myocardial mass, stroke volume and ejection fraction had no statistical difference compared to the same values derived from manual segmentation with p values all greater than 0.05. For the 475 unseen patient exams, the pipeline achieved 68% adequate segmentation in both systole and diastole, 26% needed minor adjustments in either systole or diastole, 5% needed major adjustments, and the cropping model only failed in 0.4%. Conclusion: Deep learning pipeline can provide standardised 'core-lab' segmentation for Fontan patients. This pipeline can now be applied to the >4500 cardiac magnetic resonance exams currently in the FORCE registry as well as any new patients that are recruited. △ Less

Submitted 21 March, 2023; originally announced March 2023.

Comments: 17 pages, 6 figures

Improved Segmentation of Deep Sulci in Cortical Gray Matter Using a Deep Learning Framework Incorporating Laplace's Equation

Authors: Sadhana Ravikumar, Ranjit Ittyerah, Sydney Lim, Long Xie, Sandhitsu Das, Pulkit Khandelwal, Laura E. M. Wisse, Madigan L. Bedard, John L. Robinson, Terry Schuck, Murray Grossman, John Q. Trojanowski, Edward B. Lee, M. Dylan Tisdall, Karthik Prabhakaran, John A. Detre, David J. Irwin, Winifred Trotman, Gabor Mizsei, Emilio Artacho-Pérula, Maria Mercedes Iñiguez de Onzono Martin, Maria del Mar Arroyo Jiménez, Monica Muñoz, Francisco Javier Molina Romero, Maria del Pilar Marcos Rabal , et al. (7 additional authors not shown)

Abstract: When developing tools for automated cortical segmentation, the ability to produce topologically correct segmentations is important in order to compute geometrically valid morphometry measures. In practice, accurate cortical segmentation is challenged by image artifacts and the highly convoluted anatomy of the cortex itself. To address this, we propose a novel deep learning-based cortical segmentat… ▽ More When developing tools for automated cortical segmentation, the ability to produce topologically correct segmentations is important in order to compute geometrically valid morphometry measures. In practice, accurate cortical segmentation is challenged by image artifacts and the highly convoluted anatomy of the cortex itself. To address this, we propose a novel deep learning-based cortical segmentation method in which prior knowledge about the geometry of the cortex is incorporated into the network during the training process. We design a loss function which uses the theory of Laplace's equation applied to the cortex to locally penalize unresolved boundaries between tightly folded sulci. Using an ex vivo MRI dataset of human medial temporal lobe specimens, we demonstrate that our approach outperforms baseline segmentation networks, both quantitatively and qualitatively. △ Less

Submitted 3 March, 2023; v1 submitted 1 March, 2023; originally announced March 2023.

Comments: Accepted at the 28th biennial international conference on Information Processing in Medical Imaging (IPMI 2023)

A deep learning approach to using wearable seismocardiography (SCG) for diagnosing aortic valve stenosis and predicting aortic hemodynamics obtained by 4D flow MRI

Authors: Mahmoud E. Khani, Ethan M. I. Johnson, Aparna Sodhi, Joshua Robinson, Cynthia K. Rigsby, Bradly D. Allen, Michael Markl

Abstract: In this paper, we explored the use of deep learning for the prediction of aortic flow metrics obtained using 4D flow MRI using wearable seismocardiography (SCG) devices. 4D flow MRI provides a comprehensive assessment of cardiovascular hemodynamics, but it is costly and time-consuming. We hypothesized that deep learning could be used to identify pathological changes in blood flow, such as elevated… ▽ More In this paper, we explored the use of deep learning for the prediction of aortic flow metrics obtained using 4D flow MRI using wearable seismocardiography (SCG) devices. 4D flow MRI provides a comprehensive assessment of cardiovascular hemodynamics, but it is costly and time-consuming. We hypothesized that deep learning could be used to identify pathological changes in blood flow, such as elevated peak systolic velocity Vmax in patients with heart valve diseases, from SCG signals. We also investigated the ability of this deep learning technique to differentiate between patients diagnosed with aortic valve stenosis (AS), non-AS patients with a bicuspid aortic valve (BAV), non-AS patients with a mechanical aortic valve (MAV), and healthy subjects with a normal tricuspid aortic valve (TAV). In a study of 77 subjects who underwent same-day 4D flow MRI and SCG, we found that the Vmax values obtained using deep learning and SCGs were in good agreement with those obtained by 4D flow MRI. Additionally, subjects with TAV, BAV, MAV, and AS could be classified with ROC-AUC values of 92%, 95%, 81%, and 83%, respectively. This suggests that SCG obtained using low-cost wearable electronics may be used as a supplement to 4D flow MRI exams or as a screening tool for aortic valve disease. △ Less

Submitted 5 January, 2023; originally announced January 2023.

Comments: 16 pages, 4 figures

Online Handbook of Argumentation for AI: Volume 3

Authors: Lars Bengel, Elfia Bezou-Vrakatseli, Lydia Blümel, Federico Castagna, Giulia D'Agostino, Daphne Odekerken, Minal Suresh Patil, Jordan Robinson, Hao Wu, Andreas Xydis

Abstract: This volume contains revised versions of the papers selected for the third volume of the Online Handbook of Argumentation for AI (OHAAI). Previously, formal theories of argument and argument interaction have been proposed and studied, and this has led to the more recent study of computational models of argument. Argumentation, as a field within artificial intelligence (AI), is highly relevant for… ▽ More This volume contains revised versions of the papers selected for the third volume of the Online Handbook of Argumentation for AI (OHAAI). Previously, formal theories of argument and argument interaction have been proposed and studied, and this has led to the more recent study of computational models of argument. Argumentation, as a field within artificial intelligence (AI), is highly relevant for researchers interested in symbolic representations of knowledge and defeasible reasoning. The purpose of this handbook is to provide an open access and curated anthology for the argumentation research community. OHAAI is designed to serve as a research hub to keep track of the latest and upcoming PhD-driven research on the theory and application of argumentation in all areas related to AI. △ Less

Submitted 15 December, 2022; originally announced December 2022.

A simple, efficient and scalable contrastive masked autoencoder for learning visual representations

Authors: Shlok Mishra, Joshua Robinson, Huiwen Chang, David Jacobs, Aaron Sarna, Aaron Maschinot, Dilip Krishnan

Abstract: We introduce CAN, a simple, efficient and scalable method for self-supervised learning of visual representations. Our framework is a minimal and conceptually clean synthesis of (C) contrastive learning, (A) masked autoencoders, and (N) the noise prediction approach used in diffusion models. The learning mechanisms are complementary to one another: contrastive learning shapes the embedding space ac… ▽ More We introduce CAN, a simple, efficient and scalable method for self-supervised learning of visual representations. Our framework is a minimal and conceptually clean synthesis of (C) contrastive learning, (A) masked autoencoders, and (N) the noise prediction approach used in diffusion models. The learning mechanisms are complementary to one another: contrastive learning shapes the embedding space across a batch of image samples; masked autoencoders focus on reconstruction of the low-frequency spatial correlations in a single image sample; and noise prediction encourages the reconstruction of the high-frequency components of an image. The combined approach results in a robust, scalable and simple-to-implement algorithm. The training process is symmetric, with 50% of patches in both views being masked at random, yielding a considerable efficiency improvement over prior contrastive learning methods. Extensive empirical studies demonstrate that CAN achieves strong downstream performance under both linear and finetuning evaluations on transfer learning and robustness tasks. CAN outperforms MAE and SimCLR when pre-training on ImageNet, but is especially useful for pre-training on larger uncurated datasets such as JFT-300M: for linear probe on ImageNet, CAN achieves 75.4% compared to 73.4% for SimCLR and 64.1% for MAE. The finetuned performance on ImageNet of our ViT-L model is 86.1%, compared to 85.5% for SimCLR, and 85.4% for MAE. The overall FLOPs load of SimCLR is 70% higher than CAN for ViT-L models. △ Less

Submitted 30 October, 2022; originally announced October 2022.

Comments: Mishra and Robinson contributed equally

Leveraging Large Language Models for Multiple Choice Question Answering

Authors: Joshua Robinson, Christopher Michael Rytting, David Wingate

Abstract: While large language models (LLMs) like GPT-3 have achieved impressive results on multiple choice question answering (MCQA) tasks in the zero, one, and few-shot settings, they generally lag behind the MCQA state of the art (SOTA). MCQA tasks have traditionally been presented to LLMs like cloze tasks. An LLM is conditioned on a question (without the associated answer options) and its chosen option… ▽ More While large language models (LLMs) like GPT-3 have achieved impressive results on multiple choice question answering (MCQA) tasks in the zero, one, and few-shot settings, they generally lag behind the MCQA state of the art (SOTA). MCQA tasks have traditionally been presented to LLMs like cloze tasks. An LLM is conditioned on a question (without the associated answer options) and its chosen option is the one assigned the highest probability after normalization (for length, etc.). A more natural prompting approach is to present the question and answer options to the LLM jointly and have it output the symbol (e.g., "A") associated with its chosen answer option. This approach allows the model to explicitly compare answer options, reduces computational costs, and mitigates the effects of tokenization scheme and answer option representations on answer selection. For the natural approach to be effective, the LLM it is used with must be able to associate answer options with the symbols that represent them. The LLM needs what we term multiple choice symbol binding (MCSB) ability. This ability varies greatly by model. We show that a model with high MCSB ability performs much better with the natural approach than with the traditional approach across 20 diverse datasets and largely closes the gap with the SOTA, suggesting that the MCQA ability of LLMs has been previously underestimated. △ Less

Submitted 16 March, 2023; v1 submitted 22 October, 2022; originally announced October 2022.

Comments: Accepted for ICLR 2023

Neural Set Function Extensions: Learning with Discrete Functions in High Dimensions

Authors: Nikolaos Karalias, Joshua Robinson, Andreas Loukas, Stefanie Jegelka

Abstract: Integrating functions on discrete domains into neural networks is key to developing their capability to reason about discrete objects. But, discrete domains are (1) not naturally amenable to gradient-based optimization, and (2) incompatible with deep learning architectures that rely on representations in high-dimensional vector spaces. In this work, we address both difficulties for set functions,… ▽ More Integrating functions on discrete domains into neural networks is key to developing their capability to reason about discrete objects. But, discrete domains are (1) not naturally amenable to gradient-based optimization, and (2) incompatible with deep learning architectures that rely on representations in high-dimensional vector spaces. In this work, we address both difficulties for set functions, which capture many important discrete problems. First, we develop a framework for extending set functions onto low-dimensional continuous domains, where many extensions are naturally defined. Our framework subsumes many well-known extensions as special cases. Second, to avoid undesirable low-dimensional neural network bottlenecks, we convert low-dimensional extensions into representations in high-dimensional spaces, taking inspiration from the success of semidefinite programs for combinatorial optimization. Empirically, we observe benefits of our extensions for unsupervised neural combinatorial optimization, in particular with high-dimensional representations. △ Less

Submitted 14 November, 2022; v1 submitted 8 August, 2022; originally announced August 2022.

Comments: NeurIPS 2022

An Information-theoretic Approach to Prompt Engineering Without Ground Truth Labels

Authors: Taylor Sorensen, Joshua Robinson, Christopher Michael Rytting, Alexander Glenn Shaw, Kyle Jeffrey Rogers, Alexia Pauline Delorey, Mahmoud Khalil, Nancy Fulda, David Wingate

Abstract: Pre-trained language models derive substantial linguistic and factual knowledge from the massive corpora on which they are trained, and prompt engineering seeks to align these models to specific tasks. Unfortunately, existing prompt engineering methods require significant amounts of labeled data, access to model parameters, or both. We introduce a new method for selecting prompt templates \textit{… ▽ More Pre-trained language models derive substantial linguistic and factual knowledge from the massive corpora on which they are trained, and prompt engineering seeks to align these models to specific tasks. Unfortunately, existing prompt engineering methods require significant amounts of labeled data, access to model parameters, or both. We introduce a new method for selecting prompt templates \textit{without labeled examples} and \textit{without direct access to the model}. Specifically, over a set of candidate templates, we choose the template that maximizes the mutual information between the input and the corresponding model output. Across 8 datasets representing 7 distinct NLP tasks, we show that when a template has high mutual information, it also has high accuracy on the task. On the largest model, selecting prompts with our method gets 90\% of the way from the average prompt accuracy to the best prompt accuracy and requires no ground truth labels. △ Less

Submitted 21 March, 2022; originally announced March 2022.

Sign and Basis Invariant Networks for Spectral Graph Representation Learning

Authors: Derek Lim, Joshua Robinson, Lingxiao Zhao, Tess Smidt, Suvrit Sra, Haggai Maron, Stefanie Jegelka

Abstract: We introduce SignNet and BasisNet -- new neural architectures that are invariant to two key symmetries displayed by eigenvectors: (i) sign flips, since if $v$ is an eigenvector then so is $-v$; and (ii) more general basis symmetries, which occur in higher dimensional eigenspaces with infinitely many choices of basis eigenvectors. We prove that under certain conditions our networks are universal, i… ▽ More We introduce SignNet and BasisNet -- new neural architectures that are invariant to two key symmetries displayed by eigenvectors: (i) sign flips, since if $v$ is an eigenvector then so is $-v$; and (ii) more general basis symmetries, which occur in higher dimensional eigenspaces with infinitely many choices of basis eigenvectors. We prove that under certain conditions our networks are universal, i.e., they can approximate any continuous function of eigenvectors with the desired invariances. When used with Laplacian eigenvectors, our networks are provably more expressive than existing spectral methods on graphs; for instance, they subsume all spectral graph convolutions, certain spectral graph invariants, and previously proposed graph positional encodings as special cases. Experiments show that our networks significantly outperform existing baselines on molecular graph regression, learning expressive graph representations, and learning neural fields on triangle meshes. Our code is available at https://github.com/cptq/SignNet-BasisNet . △ Less

Submitted 30 September, 2022; v1 submitted 25 February, 2022; originally announced February 2022.

Comments: 42 pages

The evolution of scientific literature as metastable knowledge states

Authors: Sai Dileep Koneru, David Rench McCauley, Michael C. Smith, David Guarrera, Jenn Robinson, Sarah Rajtmajer

Abstract: The problem of identifying common concepts in the sciences and deciding when new ideas have emerged is an open one. Metascience researchers have sought to formalize principles underlying stages in the life-cycle of scientific research, determine how knowledge is transferred between scientists and stakeholders, and understand how new ideas are generated and take hold. Here, we model the state of sc… ▽ More The problem of identifying common concepts in the sciences and deciding when new ideas have emerged is an open one. Metascience researchers have sought to formalize principles underlying stages in the life-cycle of scientific research, determine how knowledge is transferred between scientists and stakeholders, and understand how new ideas are generated and take hold. Here, we model the state of scientific knowledge immediately preceding new directions of research as a metastable state and the creation of new concepts as combinatorial innovation. We find that, through the combined use of natural language clustering and citation graph analysis, we can predict the evolution of ideas over time and thus connect a single scientific article to past and future concepts in a way that goes beyond traditional citation and reference connections. △ Less

Submitted 11 September, 2022; v1 submitted 25 February, 2022; originally announced February 2022.

The 5th Recognizing Families in the Wild Data Challenge: Predicting Kinship from Faces

Authors: Joseph P. Robinson, Can Qin, Ming Shao, Matthew A. Turk, Rama Chellappa, Yun Fu

Abstract: Recognizing Families In the Wild (RFIW), held as a data challenge in conjunction with the 16th IEEE International Conference on Automatic Face and Gesture Recognition (FG), is a large-scale, multi-track visual kinship recognition evaluation. For the fifth edition of RFIW, we continue to attract scholars, bring together professionals, publish new work, and discuss prospects. In this paper, we summa… ▽ More Recognizing Families In the Wild (RFIW), held as a data challenge in conjunction with the 16th IEEE International Conference on Automatic Face and Gesture Recognition (FG), is a large-scale, multi-track visual kinship recognition evaluation. For the fifth edition of RFIW, we continue to attract scholars, bring together professionals, publish new work, and discuss prospects. In this paper, we summarize submissions for the three tasks of this year's RFIW: specifically, we review the results for kinship verification, tri-subject verification, and family member search and retrieval. We look at the RFIW problem, share current efforts, and make recommendations for promising future directions. △ Less

Submitted 26 November, 2021; v1 submitted 31 October, 2021; originally announced November 2021.

Comments: 2021 IEEE Conference on Automatic Face and Gesture Recognition

Gray Matter Segmentation in Ultra High Resolution 7 Tesla ex vivo T2w MRI of Human Brain Hemispheres

Authors: Pulkit Khandelwal, Shokufeh Sadaghiani, Michael Tran Duong, Sadhana Ravikumar, Sydney Lim, Sanaz Arezoumandan, Claire Peterson, Eunice Chung, Madigan Bedard, Noah Capp, Ranjit Ittyerah, Elyse Migdal, Grace Choi, Emily Kopp, Bridget Loja, Eusha Hasan, Jiacheng Li, Karthik Prabhakaran, Gabor Mizsei, Marianna Gabrielyan, Theresa Schuck, John Robinson, Daniel Ohm, Edward Lee, John Q. Trojanowski , et al. (8 additional authors not shown)

Abstract: Ex vivo MRI of the brain provides remarkable advantages over in vivo MRI for visualizing and characterizing detailed neuroanatomy. However, automated cortical segmentation methods in ex vivo MRI are not well developed, primarily due to limited availability of labeled datasets, and heterogeneity in scanner hardware and acquisition protocols. In this work, we present a high resolution 7 Tesla datase… ▽ More Ex vivo MRI of the brain provides remarkable advantages over in vivo MRI for visualizing and characterizing detailed neuroanatomy. However, automated cortical segmentation methods in ex vivo MRI are not well developed, primarily due to limited availability of labeled datasets, and heterogeneity in scanner hardware and acquisition protocols. In this work, we present a high resolution 7 Tesla dataset of 32 ex vivo human brain specimens. We benchmark the cortical mantle segmentation performance of nine neural network architectures, trained and evaluated using manually-segmented 3D patches sampled from specific cortical regions, and show excellent generalizing capabilities across whole brain hemispheres in different specimens, and also on unseen images acquired at different magnetic field strength and imaging sequences. Finally, we provide cortical thickness measurements across key regions in 3D ex vivo human brain images. Our code and processed datasets are publicly available at https://github.com/Pulkit-Khandelwal/picsl-ex-vivo-segmentation. △ Less

Submitted 3 March, 2022; v1 submitted 14 October, 2021; originally announced October 2021.

Comments: Ex vivo analysis framework (work in progress 2022 at the University of Pennsylvania)

Post-Quantum Security for Ultra-Reliable Low-Latency Heterogeneous Networks

Authors: Rafael G. L. D'Oliveira, Alejandro Cohen, John Robinson, Thomas Stahlbuhk, Muriel Médard

Abstract: We consider the problem of post-quantum secure and ultra-reliable communication through a heterogeneous network consisting of multiple connections. Three performance metrics are considered: security, throughput, and in-order delivery delay. In this setting, previous work has looked, individually, at the trade-offs between in-order delivery delay and throughput, and between security and throughput.… ▽ More We consider the problem of post-quantum secure and ultra-reliable communication through a heterogeneous network consisting of multiple connections. Three performance metrics are considered: security, throughput, and in-order delivery delay. In this setting, previous work has looked, individually, at the trade-offs between in-order delivery delay and throughput, and between security and throughput. This is the first work considering the trade-off between all three for heterogeneous communication networks, while taking the computational complexity into account. We present LL-HUNCC, a low latency hybrid universal network coding cryptosystem. LL-HUNCC is an efficient coding scheme which allows for secure communications over a noisy untrusted heterogeneous network by encrypting only a small part of the information being sent. This scheme provides post-quantum security with high throughput and low in-order delivery delay guarantees. We evaluate LL-HUNCC via simulations on a setting inspired by a practical scenario for heterogeneous communications involving a satellite communication link and a 5G communication network. Under this scenario, we compare LL-HUNCC to the state-of-the-art where all communication paths are encrypted via a post-quantum public-key cryptosystem. △ Less

Submitted 13 August, 2021; originally announced August 2021.

Improved Regret Bounds for Tracking Experts with Memory

Authors: James Robinson, Mark Herbster

Abstract: We address the problem of sequential prediction with expert advice in a non-stationary environment with long-term memory guarantees in the sense of Bousquet and Warmuth [4]. We give a linear-time algorithm that improves on the best known regret bounds [26]. This algorithm incorporates a relative entropy projection step. This projection is advantageous over previous weight-sharing approaches in tha… ▽ More We address the problem of sequential prediction with expert advice in a non-stationary environment with long-term memory guarantees in the sense of Bousquet and Warmuth [4]. We give a linear-time algorithm that improves on the best known regret bounds [26]. This algorithm incorporates a relative entropy projection step. This projection is advantageous over previous weight-sharing approaches in that weight updates may come with implicit costs as in for example portfolio optimization. We give an algorithm to compute this projection step in linear time, which may be of independent interest. △ Less

Submitted 24 June, 2021; originally announced June 2021.

Can contrastive learning avoid shortcut solutions?

Authors: Joshua Robinson, Li Sun, Ke Yu, Kayhan Batmanghelich, Stefanie Jegelka, Suvrit Sra

Abstract: The generalization of representations learned via contrastive learning depends crucially on what features of the data are extracted. However, we observe that the contrastive loss does not always sufficiently guide which features are extracted, a behavior that can negatively impact the performance on downstream tasks via "shortcuts", i.e., by inadvertently suppressing important predictive features.… ▽ More The generalization of representations learned via contrastive learning depends crucially on what features of the data are extracted. However, we observe that the contrastive loss does not always sufficiently guide which features are extracted, a behavior that can negatively impact the performance on downstream tasks via "shortcuts", i.e., by inadvertently suppressing important predictive features. We find that feature extraction is influenced by the difficulty of the so-called instance discrimination task (i.e., the task of discriminating pairs of similar points from pairs of dissimilar ones). Although harder pairs improve the representation of some features, the improvement comes at the cost of suppressing previously well represented features. In response, we propose implicit feature modification (IFM), a method for altering positive and negative samples in order to guide contrastive models towards capturing a wider variety of predictive features. Empirically, we observe that IFM reduces feature suppression, and as a result improves performance on vision and medical imaging tasks. The code is available at: \url{https://github.com/joshr17/IFM}. △ Less

Submitted 19 December, 2021; v1 submitted 21 June, 2021; originally announced June 2021.

Comments: NeurIPS 2021

A Matrix Autoencoder Framework to Align the Functional and Structural Connectivity Manifolds as Guided by Behavioral Phenotypes

Authors: Niharika Shimona D'Souza, Mary Beth Nebel, Deana Crocetti, Nicholas Wymbs, Joshua Robinson, Stewart Mostofsky, Archana Venkataraman

Abstract: We propose a novel matrix autoencoder to map functional connectomes from resting state fMRI (rs-fMRI) to structural connectomes from Diffusion Tensor Imaging (DTI), as guided by subject-level phenotypic measures. Our specialized autoencoder infers a low dimensional manifold embedding for the rs-fMRI correlation matrices that mimics a canonical outer-product decomposition. The embedding is simultan… ▽ More We propose a novel matrix autoencoder to map functional connectomes from resting state fMRI (rs-fMRI) to structural connectomes from Diffusion Tensor Imaging (DTI), as guided by subject-level phenotypic measures. Our specialized autoencoder infers a low dimensional manifold embedding for the rs-fMRI correlation matrices that mimics a canonical outer-product decomposition. The embedding is simultaneously used to reconstruct DTI tractography matrices via a second manifold alignment decoder and to predict inter-subject phenotypic variability via an artificial neural network. We validate our framework on a dataset of 275 healthy individuals from the Human Connectome Project database and on a second clinical dataset consisting of 57 subjects with Autism Spectrum Disorder. We demonstrate that the model reliably recovers structural connectivity patterns across individuals, while robustly extracting predictive and interpretable brain biomarkers in a cross-validated setting. Finally, our framework outperforms several baselines at predicting behavioral phenotypes in both real-world datasets. △ Less

Submitted 9 July, 2021; v1 submitted 29 May, 2021; originally announced May 2021.

Balancing Biases and Preserving Privacy on Balanced Faces in the Wild

Authors: Joseph P Robinson, Can Qin, Yann Henon, Samson Timoner, Yun Fu

Abstract: There are demographic biases present in current facial recognition (FR) models. To measure these biases across different ethnic and gender subgroups, we introduce our Balanced Faces in the Wild (BFW) dataset. This dataset allows for the characterization of FR performance per subgroup. We found that relying on a single score threshold to differentiate between genuine and imposters sample pairs lead… ▽ More There are demographic biases present in current facial recognition (FR) models. To measure these biases across different ethnic and gender subgroups, we introduce our Balanced Faces in the Wild (BFW) dataset. This dataset allows for the characterization of FR performance per subgroup. We found that relying on a single score threshold to differentiate between genuine and imposters sample pairs leads to suboptimal results. Additionally, performance within subgroups often varies significantly from the global average. Therefore, specific error rates only hold for populations that match the validation data. To mitigate imbalanced performances, we propose a novel domain adaptation learning scheme that uses facial features extracted from state-of-the-art neural networks. This scheme boosts the average performance and preserves identity information while removing demographic knowledge. Removing demographic knowledge prevents potential biases from affecting decision-making and protects privacy by eliminating demographic information. We explore the proposed method and demonstrate that subgroup classifiers can no longer learn from features projected using our domain adaptation scheme. For access to the source code and data, please visit https://github.com/visionjo/facerec-bias-bfw. △ Less

Submitted 5 July, 2023; v1 submitted 16 March, 2021; originally announced March 2021.

Comments: arXiv admin note: text overlap with arXiv:2102.08941

Automatic Face Understanding: Recognizing Families in Photos

Authors: Joseph P Robinson

Abstract: We built the largest database for kinship recognition. The data were labeled using a novel clustering algorithm that used label proposals as side information to guide more accurate clusters. Great savings in time and human input was had. Statistically, FIW shows enormous gains over its predecessors. We have several benchmarks in kinship verification, family classification, tri-subject verification… ▽ More We built the largest database for kinship recognition. The data were labeled using a novel clustering algorithm that used label proposals as side information to guide more accurate clusters. Great savings in time and human input was had. Statistically, FIW shows enormous gains over its predecessors. We have several benchmarks in kinship verification, family classification, tri-subject verification, and large-scale search and retrieval. We also trained CNNs on FIW and deployed the model on the renowned KinWild I and II to gain SOTA. Most recently, we further augmented FIW with MM. Now, video dynamics, audio, and text captions can be used in the decision making of kinship recognition systems. We expect FIW will significantly impact research and reality. Additionally, we tackled the classic problem of facial landmark localization. A majority of these networks have objectives based on L1 or L2 norms, which inherit several disadvantages. The locations of landmarks are determined from generated heatmaps from which predicted landmark locations get penalized without accounting for the spread: a high scatter corresponds to low confidence and vice-versa. To address this, we introduced an objective that penalizes for low confidence. Another issue is a dependency on labeled data, which is expensive to collect and susceptible to error. We addressed both issues by proposing an adversarial training framework that leverages unlabeled data to improve model performance. Our method claims SOTA on renowned benchmarks. Furthermore, our model is robust with a reduced size: 1/8 the number of channels is comparable to SOTA in real-time on a CPU. Finally, we built BFW to serve as a proxy to measure bias across ethnicity and gender subgroups, allowing us to characterize FR performances per subgroup. We show performances are non-optimal when a single threshold is used to determine whether sample pairs are genuine. △ Less

Submitted 10 January, 2021; originally announced February 2021.

Comments: PhD Thesis

Multimodal In-bed Pose and Shape Estimation under the Blankets

Authors: Yu Yin, Joseph P. Robinson, Yun Fu

Abstract: Humans spend vast hours in bed -- about one-third of the lifetime on average. Besides, a human at rest is vital in many healthcare applications. Typically, humans are covered by a blanket when resting, for which we propose a multimodal approach to uncover the subjects so their bodies at rest can be viewed without the occlusion of the blankets above. We propose a pyramid scheme to effectively fuse… ▽ More Humans spend vast hours in bed -- about one-third of the lifetime on average. Besides, a human at rest is vital in many healthcare applications. Typically, humans are covered by a blanket when resting, for which we propose a multimodal approach to uncover the subjects so their bodies at rest can be viewed without the occlusion of the blankets above. We propose a pyramid scheme to effectively fuse the different modalities in a way that best leverages the knowledge captured by the multimodal sensors. Specifically, the two most informative modalities (i.e., depth and infrared images) are first fused to generate good initial pose and shape estimation. Then pressure map and RGB images are further fused one by one to refine the result by providing occlusion-invariant information for the covered part, and accurate shape information for the uncovered part, respectively. However, even with multimodal data, the task of detecting human bodies at rest is still very challenging due to the extreme occlusion of bodies. To further reduce the negative effects of the occlusion from blankets, we employ an attention-based reconstruction module to generate uncovered modalities, which are further fused to update current estimation via a cyclic fashion. Extensive experiments validate the superiority of the proposed model over others. △ Less

Submitted 12 December, 2020; originally announced December 2020.

SuperFront: From Low-resolution to High-resolution Frontal Face Synthesis

Authors: Yu Yin, Joseph P. Robinson, Songyao Jiang, Yue Bai, Can Qin, Yun Fu

Abstract: Advances in face rotation, along with other face-based generative tasks, are more frequent as we advance further in topics of deep learning. Even as impressive milestones are achieved in synthesizing faces, the importance of preserving identity is needed in practice and should not be overlooked. Also, the difficulty should not be more for data with obscured faces, heavier poses, and lower quality.… ▽ More Advances in face rotation, along with other face-based generative tasks, are more frequent as we advance further in topics of deep learning. Even as impressive milestones are achieved in synthesizing faces, the importance of preserving identity is needed in practice and should not be overlooked. Also, the difficulty should not be more for data with obscured faces, heavier poses, and lower quality. Existing methods tend to focus on samples with variation in pose, but with the assumption data is high in quality. We propose a generative adversarial network (GAN) -based model to generate high-quality, identity preserving frontal faces from one or multiple low-resolution (LR) faces with extreme poses. Specifically, we propose SuperFront-GAN (SF-GAN) to synthesize a high-resolution (HR), frontal face from one-to-many LR faces with various poses and with the identity-preserved. We integrate a super-resolution (SR) side-view module into SF-GAN to preserve identity information and fine details of the side-views in HR space, which helps model reconstruct high-frequency information of faces (i.e., periocular, nose, and mouth regions). Moreover, SF-GAN accepts multiple LR faces as input, and improves each added sample. We squeeze additional gain in performance with an orthogonal constraint in the generator to penalize redundant latent representations and, hence, diversify the learned features space. Quantitative and qualitative results demonstrate the superiority of SF-GAN over others. △ Less

Submitted 7 December, 2020; originally announced December 2020.

Recommendations for Bayesian hierarchical model specifications for case-control studies in mental health

Authors: Vincent Valton, Toby Wise, Oliver J. Robinson

Abstract: Hierarchical model fitting has become commonplace for case-control studies of cognition and behaviour in mental health. However, these techniques require us to formalise assumptions about the data-generating process at the group level, which may not be known. Specifically, researchers typically must choose whether to assume all subjects are drawn from a common population, or to model them as deriv… ▽ More Hierarchical model fitting has become commonplace for case-control studies of cognition and behaviour in mental health. However, these techniques require us to formalise assumptions about the data-generating process at the group level, which may not be known. Specifically, researchers typically must choose whether to assume all subjects are drawn from a common population, or to model them as deriving from separate populations. These assumptions have profound implications for computational psychiatry, as they affect the resulting inference (latent parameter recovery) and may conflate or mask true group-level differences. To test these assumptions we ran systematic simulations on synthetic multi-group behavioural data from a commonly used multi-armed bandit task (reinforcement learning task). We then examined recovery of group differences in latent parameter space under the two commonly used generative modelling assumptions: (1) modelling groups under a common shared group-level prior (assuming all participants are generated from a common distribution, and are likely to share common characteristics); (2) modelling separate groups based on symptomatology or diagnostic labels, resulting in separate group-level priors. We evaluated the robustness of these approaches to variations in data quality and prior specifications on a variety of metrics. We found that fitting groups separately (assumptions 2), provided the most accurate and robust inference across all conditions. Our results suggest that when dealing with data from multiple clinical groups, researchers should analyse patient and control groups separately as it provides the most accurate and robust recovery of the parameters of interest. △ Less

Submitted 3 November, 2020; originally announced November 2020.

Comments: Machine Learning for Health (ML4H) at NeurIPS 2020 - Extended Abstract

Contrastive Learning with Hard Negative Samples

Authors: Joshua Robinson, Ching-Yao Chuang, Suvrit Sra, Stefanie Jegelka

Abstract: How can you sample good negative examples for contrastive learning? We argue that, as with metric learning, contrastive learning of representations benefits from hard negative samples (i.e., points that are difficult to distinguish from an anchor point). The key challenge toward using hard negatives is that contrastive methods must remain unsupervised, making it infeasible to adopt existing negati… ▽ More How can you sample good negative examples for contrastive learning? We argue that, as with metric learning, contrastive learning of representations benefits from hard negative samples (i.e., points that are difficult to distinguish from an anchor point). The key challenge toward using hard negatives is that contrastive methods must remain unsupervised, making it infeasible to adopt existing negative sampling strategies that use true similarity information. In response, we develop a new family of unsupervised sampling methods for selecting hard negative samples where the user can control the hardness. A limiting case of this sampling results in a representation that tightly clusters each class, and pushes different classes as far apart as possible. The proposed method improves downstream performance across multiple modalities, requires only few additional lines of code to implement, and introduces no computational overhead. △ Less

Submitted 24 January, 2021; v1 submitted 9 October, 2020; originally announced October 2020.

Comments: Published as a conference paper at ICLR 2021

Deep sr-DDL: Deep Structurally Regularized Dynamic Dictionary Learning to Integrate Multimodal and Dynamic Functional Connectomics data for Multidimensional Clinical Characterizations

Authors: Niharika Shimona D'Souza, Mary Beth Nebel, Deana Crocetti, Nicholas Wymbs, Joshua Robinson, Stewart H. Mostofsky, Archana Venkataraman

Abstract: We propose a novel integrated framework that jointly models complementary information from resting-state functional MRI (rs-fMRI) connectivity and diffusion tensor imaging (DTI) tractography to extract biomarkers of brain connectivity predictive of behavior. Our framework couples a generative model of the connectomics data with a deep network that predicts behavioral scores. The generative compone… ▽ More We propose a novel integrated framework that jointly models complementary information from resting-state functional MRI (rs-fMRI) connectivity and diffusion tensor imaging (DTI) tractography to extract biomarkers of brain connectivity predictive of behavior. Our framework couples a generative model of the connectomics data with a deep network that predicts behavioral scores. The generative component is a structurally-regularized Dynamic Dictionary Learning (sr-DDL) model that decomposes the dynamic rs-fMRI correlation matrices into a collection of shared basis networks and time varying subject-specific loadings. We use the DTI tractography to regularize this matrix factorization and learn anatomically informed functional connectivity profiles. The deep component of our framework is an LSTM-ANN block, which uses the temporal evolution of the subject-specific sr-DDL loadings to predict multidimensional clinical characterizations. Our joint optimization strategy collectively estimates the basis networks, the subject-specific time-varying loadings, and the neural network weights. We validate our framework on a dataset of neurotypical individuals from the Human Connectome Project (HCP) database to map to cognition and on a separate multi-score prediction task on individuals diagnosed with Autism Spectrum Disorder (ASD) in a five-fold cross validation setting. Our hybrid model outperforms several state-of-the-art approaches at clinical outcome prediction and learns interpretable multimodal neural signatures of brain organization. △ Less

Submitted 21 November, 2024; v1 submitted 27 August, 2020; originally announced August 2020.

Families In Wild Multimedia: A Multimodal Database for Recognizing Kinship

Authors: Joseph P. Robinson, Zaid Khan, Yu Yin, Ming Shao, Yun Fu

Abstract: Kinship, a soft biometric detectable in media, is fundamental for a myriad of use-cases. Despite the difficulty of detecting kinship, annual data challenges using still-images have consistently improved performances and attracted new researchers. Now, systems reach performance levels unforeseeable a decade ago, closing in on performances acceptable to deploy in practice. Like other biometric tasks… ▽ More Kinship, a soft biometric detectable in media, is fundamental for a myriad of use-cases. Despite the difficulty of detecting kinship, annual data challenges using still-images have consistently improved performances and attracted new researchers. Now, systems reach performance levels unforeseeable a decade ago, closing in on performances acceptable to deploy in practice. Like other biometric tasks, we expect systems can receive help from other modalities. We hypothesize that adding modalities to FIW, which has only still-images, will improve performance. Thus, to narrow the gap between research and reality and enhance the power of kinship recognition systems, we extend FIW with multimedia (MM) data (i.e., video, audio, and text captions). Specifically, we introduce the first publicly available multi-task MM kinship dataset. To build FIW MM, we developed machinery to automatically collect, annotate, and prepare the data, requiring minimal human input and no financial cost. The proposed MM corpus allows the problem statements to be more realistic template-based protocols. We show significant improvements in all benchmarks with the added modalities. The results highlight edge cases to inspire future research with different areas of improvement. FIW MM supplies the data needed to increase the potential of automated systems to detect kinship in MM. It also allows experts from diverse fields to collaborate in novel ways. △ Less

Submitted 1 October, 2021; v1 submitted 28 July, 2020; originally announced July 2020.

Journal ref: IEEE Transactions on Multimedia (2021)

A Deep-Generative Hybrid Model to Integrate Multimodal and Dynamic Connectivity for Predicting Spectrum-Level Deficits in Autism

Authors: Niharika Shimona D'Souza, Mary Beth Nebel, Deana Crocetti, Nicholas Wymbs, Joshua Robinson, Stewart Mostofsky, Archana Venkataraman

Abstract: We propose an integrated deep-generative framework, that jointly models complementary information from resting-state functional MRI (rs-fMRI) connectivity and diffusion tensor imaging (DTI) tractography to extract predictive biomarkers of a disease. The generative part of our framework is a structurally-regularized Dynamic Dictionary Learning (sr-DDL) model that decomposes the dynamic rs-fMRI corr… ▽ More We propose an integrated deep-generative framework, that jointly models complementary information from resting-state functional MRI (rs-fMRI) connectivity and diffusion tensor imaging (DTI) tractography to extract predictive biomarkers of a disease. The generative part of our framework is a structurally-regularized Dynamic Dictionary Learning (sr-DDL) model that decomposes the dynamic rs-fMRI correlation matrices into a collection of shared basis networks and time varying patient-specific loadings. This matrix factorization is guided by the DTI tractography matrices to learn anatomically informed connectivity profiles. The deep part of our framework is an LSTM-ANN block, which models the temporal evolution of the patient sr-DDL loadings to predict multidimensional clinical severity. Our coupled optimization procedure collectively estimates the basis networks, the patient-specific dynamic loadings, and the neural network weights. We validate our framework on a multi-score prediction task in 57 patients diagnosed with Autism Spectrum Disorder (ASD). Our hybrid model outperforms state-of-the-art baselines in a five-fold cross validated setting and extracts interpretable multimodal neural signatures of brain dysfunction in ASD. △ Less

Submitted 21 November, 2024; v1 submitted 3 July, 2020; originally announced July 2020.

Debiased Contrastive Learning

Authors: Ching-Yao Chuang, Joshua Robinson, Lin Yen-Chen, Antonio Torralba, Stefanie Jegelka

Abstract: A prominent technique for self-supervised representation learning has been to contrast semantically similar and dissimilar pairs of samples. Without access to labels, dissimilar (negative) points are typically taken to be randomly sampled datapoints, implicitly accepting that these points may, in reality, actually have the same label. Perhaps unsurprisingly, we observe that sampling negative examp… ▽ More A prominent technique for self-supervised representation learning has been to contrast semantically similar and dissimilar pairs of samples. Without access to labels, dissimilar (negative) points are typically taken to be randomly sampled datapoints, implicitly accepting that these points may, in reality, actually have the same label. Perhaps unsurprisingly, we observe that sampling negative examples from truly different labels improves performance, in a synthetic setting where labels are available. Motivated by this observation, we develop a debiased contrastive objective that corrects for the sampling of same-label datapoints, even without knowledge of the true labels. Empirically, the proposed objective consistently outperforms the state-of-the-art for representation learning in vision, language, and reinforcement learning benchmarks. Theoretically, we establish generalization bounds for the downstream classification task. △ Less

Submitted 21 October, 2020; v1 submitted 1 July, 2020; originally announced July 2020.

Journal ref: Advances in Neural Information Processing Systems (2020)

Survey on the Analysis and Modeling of Visual Kinship: A Decade in the Making

Authors: Joseph P Robinson, Ming Shao, Yun Fu

Abstract: Kinship recognition is a challenging problem with many practical applications. With much progress and milestones having been reached after ten years - we are now able to survey the research and create new milestones. We review the public resources and data challenges that enabled and inspired many to hone-in on the views of automatic kinship recognition in the visual domain. The different tasks ar… ▽ More Kinship recognition is a challenging problem with many practical applications. With much progress and milestones having been reached after ten years - we are now able to survey the research and create new milestones. We review the public resources and data challenges that enabled and inspired many to hone-in on the views of automatic kinship recognition in the visual domain. The different tasks are described in technical terms and syntax consistent across the problem domain and the practical value of each discussed and measured. State-of-the-art methods for visual kinship recognition problems, whether to discriminate between or generate from, are examined. As part of such, we review systems proposed as part of a recent data challenge held in conjunction with the 2020 IEEE Conference on Automatic Face and Gesture Recognition. We establish a stronghold for the state of progress for the different problems in a consistent manner. This survey will serve as the central resource for the work of the next decade to build upon. For the tenth anniversary, the demo code is provided for the various kin-based tasks. Detecting relatives with visual recognition and classifying the relationship is an area with high potential for impact in research and practice.IEEE Transactions on pattern analysis and machine intelligence △ Less

Submitted 23 February, 2021; v1 submitted 29 June, 2020; originally announced June 2020.

Journal ref: IEEE Transactions on pattern analysis and machine intelligence (2021)