AIConfPaper
← Search

Sanmi Koyejo

77 accepted papers

2026

AbdCTBench: Learning Clinical Biomarker Representations from Abdominal Surface Geometry

Muhammad Ahmed Chaudhry, Suhana Bedi, Pola Lydia Lagari, Brian T Layden, William Galanter, Ayis Pyrros +1
ICLR 2026poster

Body composition analysis through CT and MRI imaging provides critical insights for cardio-metabolic health assessment but remains limited by accessibility barriers including radiation exposure, high costs, and infrastructure requirements. We present AbdCTBench, a large-scale dataset containing 23,5…

Cited by 0Source ↗Scholar ↗
2026

Confidence and Difficulty-Adaptive Policy Optimization for LLM Reasoning

(Andrew) Zhanke Zhou, Xiangyu Lu, Chentao Cao, Brando Miranda, Tongliang Liu, Bo Han +1
ICML 2026poster

RL with verifiable rewards can substantially improve LLM reasoning, yet standard GRPO-style training often uses uniform sampling and near-uniform weighting, leading to inefficient computation allocation. We study GRPO by tracking token log-probabilities, group-normalized advantages, and induced toke…

Cited by 0Source ↗Scholar ↗
2026

Gaming Consensus: Coordinated Manipulation in Crowdsourced Fact-Checking

Nikil Selvam, Jay Baxter, Sophie Hilgard, Brad Miller, Keith Coleman, Ellen Vitercik +1
ICML 2026poster

Crowdsourced fact-checking systems have been widely adopted by major social media companies such as X, Meta, Tiktok, and Google with the aim of combating misinformation at scale without relying on centralized editorial control. These systems have been developed around a common underlying algorithm: …

Cited by 0Source ↗Scholar ↗
2026

Item Response Scaling Laws: A Measurement Theory Approach for Efficient and Generalizable Neural Scaling Estimation

Sang Truong, Yuheng Tu, Rylan Schaeffer, Sanmi Koyejo
ICML 2026poster

Scaling laws provide a fundamental framework for understanding the performance of Large Language Models (LLMs), yet deriving them requires prohibitively expensive evaluations across thousands of checkpoints or millions of inference samples. To address this, we introduce Item Response Scaling Laws (I…

Cited by 0Source ↗Scholar ↗
2026

Landscape of Thoughts: Visualizing the Reasoning Process of Large Language Models

Zhanke Zhou, Zhaocheng Zhu, Xuan Li, Mikhail Galkin, Xiao Feng, Sanmi Koyejo +2
ICLR 2026poster

Numerous applications of large language models (LLMs) rely on their ability to perform step-by-step reasoning. However, the reasoning behavior of LLMs remains poorly understood, posing challenges to research, development, and safety. To address this gap, we introduce landscape of thoughts (LoT), the…

Cited by 0Source ↗code ↗Scholar ↗
2026

No, of Course I Can! Deeper Fine-Tuning Attacks That Bypass Token-Level Safety Mechanisms

Joshua Kazdan, Abhay Puri, Rylan Schaeffer, Lisa Yu, Chris Cundy, Jason Stanley +2
ICLR 2026poster

Leading language model (LM) providers like OpenAI and Anthopic allow customers to fine-tune frontier LMs for specific use cases. To prevent abuse, these providers apply filters to block fine-tuning on overtly harmful data. In this setting, we make three contributions: First, while past work has show…

Cited by 0Source ↗Scholar ↗
2026

Noise Tectonics: Measuring the Stability of AI Benchmark Ecosystems

Michael Hardy, Sang Truong, Anka Reuel, Lijin Zhang, Jodi Casabianca, Yash Dave +3
ICML 2026poster

AI benchmark ecosystems compress rich evaluation data into aggregate leaderboard scores, but these scores contain substantial measurement noise whose sources and magnitudes remain unquantified. Without systematic methods to measure this noise and separate signal from artifact, it is unclear when ben…

Cited by 0Source ↗Scholar ↗
2026

Position: Beyond Sensitive Attributes, ML Fairness Should Quantify Structural Injustice via Social Determinants

Zeyu Tang, Alex John London, Atoosa Kasirzadeh, Sarah de Ramirez, Peter Spirtes, Kun Zhang +1
ICML 2026poster

Algorithmic fairness research has largely framed _unfairness as discrimination_ along _sensitive attributes_. However, this approach limits visibility into _unfairness as structural injustice_ instantiated through _social determinants_, which are contextual variables that shape attributes and outcom…

Cited by 0Source ↗Scholar ↗
2026

Position: Multiple Definitions & Unrealistic Assumptions of Model Collapse Distract from Real World Threats

Rylan Schaeffer, Joshua Kazdan, Alvan Arulandu, Sanmi Koyejo
ICML 2026poster

The proliferation of AI-generated content online has fueled concerns over \textit{model collapse}, a degradation in future generative models' performance when trained on synthetic data generated by earlier models. Industry leaders, premier research journals and popular science publications alike hav…

Cited by 0Source ↗Scholar ↗
2026

Position: There are futures that benchmark-driven AI cannot see

Sobhan Lotfi, Ava Iranmanesh, Lachin Naghashyar, Ali Shirali, Fateme Haredasht, Sanmi Koyejo +6
ICML 2026oral

Breakthroughs often come from ideas we could not have predicted in advance. In biology, this is called exaptation: traits evolved for one function become decisive for another. Scientific progress works similarly, but only if ideas survive periods when they appear uncompetitive by current metrics. Th…

Cited by 0Source ↗Scholar ↗
2026

Pretraining Scaling Laws for Generative Evaluations of Language Models

Rylan Schaeffer, Noam Itzhak Levi, Brando Miranda, Sanmi Koyejo
ICLR 2026poster

Neural scaling laws have driven the field's ever-expanding exponential growth in parameters, data and compute. While scaling behaviors for pretraining losses and discriminative benchmarks are well established, generative benchmarks such as mathematical problem-solving or software engineering remain…

Cited by 0Source ↗Scholar ↗
2026

Truthfulness Does Not Scale Like Reasoning: Why Polling Fails as a Proxy Verifier

Yegor Denisov-Blanch, Joshua Kazdan, Jessica Chudnovsky, Rylan Schaeffer, Sheng Guan, Soji Adeshina +1
ICML 2026poster

Pass@$k$ and other methods of scaling inference compute can improve language model performance in domains with external verifiers, including mathematics and code, where incorrect candidates can be filtered reliably. This raises a natural question: can we similarly scale compute to elicit gains in tr…

Cited by 0Source ↗Scholar ↗
2026

When AI Benchmarks Plateau: A Systematic Study of Benchmark Saturation

Mubashara Akhtar, Anka Reuel, Prajna Soni, Sanchit Ahuja, Pawan Sasanka Ammanamanchi, Ruchit Rawal +31
ICML 2026poster

Artificial Intelligence (AI) benchmarks play a central role in measuring progress in model development and guiding deployment decisions. However, many benchmarks quickly become saturated, meaning that they can no longer differentiate between the best-performing models, diminishing their long-term va…

Cited by 0Source ↗Scholar ↗
2026

Who Evaluates AI's Social Impacts? Mapping Coverage and Gaps in First and Third Party Evaluations

Anka Reuel, Avijit Ghosh, Jenny Chim, Andrew Tran, Yanan Long, Jennifer Mickel +29
ICML 2026poster

Foundation models are increasingly central to high-stakes AI systems, and governance frameworks now depend on evaluations to assess their risks and capabilities. Although general capability evaluations are widespread, social impact assessments covering bias, fairness, privacy, environmental costs, a…

Cited by 0Source ↗Scholar ↗
2025

Aligning Compound AI Systems via System-level DPO

Xiangwen Wang, Yibo Jacky Zhang, Zhoujie Ding, Katherine Tsai, Haolun Wu, Sanmi Koyejo
NeurIPS 2025poster

Compound AI systems, comprising multiple interacting components such as LLMs, foundation models, and external tools, have demonstrated remarkable improvements compared to single models in various tasks. To ensure their effective deployment in real-world applications, aligning these systems with huma…

Cited by 0Source ↗Scholar ↗
2025

AutoRedTeamer: Autonomous Red Teaming with Lifelong Attack Integration

Andy Zhou, Kevin Wu, Francesco Pinto, Zhaorun Chen, Yi Zeng, Yu Yang +4
NeurIPS 2025poster

As large language models (LLMs) become increasingly capable, security and safety evaluation are crucial. While current red teaming approaches have made strides in assessing LLM vulnerabilities, they often rely heavily on human input and lack comprehensive coverage of emerging attack vectors. This pa…

Cited by 0Source ↗Scholar ↗
2025

Certified Unlearning for Neural Networks

Anastasia Koloskova, Youssef Allouah, Animesh Jha, Rachid Guerraoui, Sanmi Koyejo
ICML 2025poster

We address the problem of machine unlearning, where the goal is to remove the influence of specific training data from a model upon request, motivated by privacy concerns and regulatory requirements such as the “right to be forgotten.” Unfortunately, existing methods rely on restrictive assumptions…

Cited by 0Source ↗PDF ↗code ↗Scholar ↗
2025

Collapse or Thrive: Perils and Promises of Synthetic Data in a Self-Generating World

Joshua Kazdan, Rylan Schaeffer, Apratim Dey, Matthias Gerstgrasser, Rafael Rafailov, David L. Donoho +1
ICML 2025poster

What happens when generative machine learning models are pretrained on web-scale datasets containing data generated by earlier models? Some prior work warns of “model collapse” as the web is overwhelmed by synthetic data; other work suggests the problem can be contained (i.e. collapse can be avoided…

Cited by 8Source ↗PDF ↗Scholar ↗
2025

Context Clues: Evaluating Long Context Models for Clinical Prediction Tasks on EHR Data

Michael Wornow, Suhana Bedi, Miguel Angel Fuentes Hernandez, Ethan Steinberg, Jason Alan Fries, Christopher Re +2
ICLR 2025poster

Foundation Models (FMs) trained on Electronic Health Records (EHRs) have achieved state-of-the-art results on numerous clinical prediction tasks. However, prior EHR FMs typically have context windows of $<$1k tokens, which prevents them from modeling full patient EHRs which can exceed 10k's of event…

Cited by 1Source ↗PDF ↗Scholar ↗
2025

Decision from Suboptimal Classifiers: Excess Risk Pre- and Post-Calibration

Alexandre Perez-Lebel, Gael Varoquaux, Sanmi Koyejo, Matthieu Doutreligne, Marine Le Morvan
AISTATS 2025poster

Probabilistic classifiers are central for making informed decisions under uncertainty. Based on the maximum expected utility principle, optimal decision rules can be derived using the posterior class probabilities and misclassification costs. Yet, in practice only learned approximations of the oracl…

Cited by 0Source ↗code ↗Scholar ↗
2025

Failures to Find Transferable Image Jailbreaks Between Vision-Language Models

Rylan Schaeffer, Dan Valentine, Luke Bailey, James Chua, Cristobal Eyzaguirre, Zane Durante +10
ICLR 2025poster

The integration of new modalities into frontier AI systems offers exciting capabilities, but also increases the possibility such systems can be adversarially manipulated in undesirable ways. In this work, we focus on a popular class of vision-language models (VLMs) that generate text outputs conditi…

Cited by 3Source ↗PDF ↗Scholar ↗
2025

Fairness through Difference Awareness: Measuring Desired Group Discrimination in LLMs

Angelina Wang, Michelle Phan, Daniel E. Ho, Sanmi Koyejo
ACL 2025long

Algorithmic fairness has conventionally adopted the mathematically convenient perspective of racial color-blindness (i.e., difference unaware treatment). However, we contend that in a range of important settings, group difference awareness matters. For example, differentiating between groups may be…

Cited by 0Source ↗PDF ↗code ↗Scholar ↗
2025

Fantastic Bugs and Where to Find Them in AI Benchmarks

Sang T. Truong, Yuheng Tu, Michael Hardy, Anka Reuel, Zeyu Tang, Jirayu Burapacheep +5
NeurIPS 2025poster

Benchmarks are pivotal in driving AI progress, and invalid benchmark questions frequently undermine their reliability. Manually identifying and correcting errors among thousands of benchmark questions is not only infeasible but also a critical bottleneck for reliable evaluation. In this work, we int…

Cited by 0Source ↗Scholar ↗
2025

From Passive to Active Reasoning: Can Large Language Models Ask the Right Questions under Incomplete Information?

Zhanke Zhou, Xiao Feng, Zhaocheng Zhu, Jiangchao Yao, Sanmi Koyejo, Bo Han
ICML 2025poster

While existing benchmarks probe the reasoning abilities of large language models (LLMs) across diverse domains, they predominantly assess passive reasoning, providing models with all the information needed to reach a solution. By contrast, active reasoning—where an LLM must interact with external sy…

Cited by 0Source ↗PDF ↗code ↗Scholar ↗
2025

How Do Large Language Monkeys Get Their Power (Laws)?

Rylan Schaeffer, Joshua Kazdan, John Hughes, Jordan Juravsky, Sara Price, Aengus Lynch +4
ICML 2025oral

Recent research across mathematical problem solving, proof assistant programming and multimodal jailbreaking documents a striking finding: when (multimodal) language model tackle a suite of tasks with multiple attempts per task -- succeeding if any attempt is correct -- then the negative log of the…

Cited by 0Source ↗PDF ↗Scholar ↗
2025

KGGen: Extracting Knowledge Graphs from Plain Text with Language Models

Belinda Mo, Kyssen Yu, Joshua Kazdan, Proud Mpala, Lisa Yu, Charilaos I. Kanatsoulis +1
NeurIPS 2025poster

Recent interest in building foundation models for knowledge graphs has highlighted a fundamental challenge: knowledge graph data is scarce. The best-known knowl- edge graphs are primarily human-labeled, created by pattern-matching, or extracted using early NLP techniques. While human-generated knowl…

Cited by 0Source ↗code ↗Scholar ↗
2025

Language Models May Verbatim Complete Text They Were Not Explicitly Trained On

Ken Liu, Christopher A. Choquette-Choo, Matthew Jagielski, Peter Kairouz, Sanmi Koyejo, Percy Liang +1
ICML 2025spotlight

An important question today is whether a given text was used to train a large language model (LLM). A completion test is often employed: check if the LLM completes a sufficiently complex text. This, however, requires a ground-truth definition of membership; most commonly, it is defined as a member b…

Cited by 0Source ↗PDF ↗Scholar ↗
2025

Learning to (Learn at Test Time): RNNs with Expressive Hidden States

Yu Sun, Xinhao Li, Karan Dalal, Jiarui Xu, Arjun Vikram, Genghan Zhang +6
ICML 2025spotlight

Self-attention performs well in long context but has quadratic complexity. Existing RNN layers have linear complexity, but their performance in long context is limited by the expressive power of their hidden states. We present a practical framework for instantiating sequence modeling layers with lin…

Cited by 90Source ↗PDF ↗code ↗Scholar ↗
2025

Logits are All We Need to Adapt Closed Models

Gaurush Hiranandani, Haolun Wu, Subhojyoti Mukherjee, Sanmi Koyejo
ICML 2025poster

Many commercial Large Language Models (LLMs) are often closed-source, limiting developers to prompt tuning for aligning content generation with specific applications. While these models currently do not provide access to token logits, we argue that if such access were available, it would enable more…

Cited by 0Source ↗PDF ↗code ↗Scholar ↗
2025

Machine Unlearning Doesn't Do What You Think: Lessons for Generative AI Policy and Research

A. Feder Cooper, Christopher A. Choquette-Choo, Miranda Bogen, Kevin Klyman, Matthew Jagielski, Katja Filippova +31
NeurIPS 2025oral

"Machine unlearning" is a popular proposed solution for mitigating the existence of content in an AI model that is problematic for legal or moral reasons, including privacy, copyright, safety, and more. For example, unlearning is often invoked as a solution for removing the effects of specific infor…

Cited by 0Source ↗Scholar ↗
2025

Position: Machine Learning Conferences Should Establish a "Refutations and Critiques" Track

Rylan Schaeffer, Joshua Kazdan, Yegor Denisov-Blanch, Brando Miranda, Matthias Gerstgrasser, Susan Zhang +8
NeurIPS 2025oral

Science progresses by iteratively advancing and correcting humanity's understanding of the world. In machine learning (ML) research, rapid advancements have led to an explosion of publications, but have also led to misleading, incorrect, flawed or perhaps even fraudulent studies being accepted and s…

Cited by 0Source ↗Scholar ↗
2025

Putnam-AXIOM: A Functional & Static Benchmark for Measuring Higher Level Mathematical Reasoning in LLMs

Aryan Gulati, Brando Miranda, Eric Chen, Emily Xia, Kai Fronsdal, Bruno de Moraes Dumont +1
ICML 2025poster

Current mathematical reasoning benchmarks for large language models (LLMs) are approaching saturation, with some achieving $>$ 90% accuracy, and are increasingly compromised by training-set contamination. We introduce Putnam-AXIOM, a benchmark of 522 university-level competition problems drawn from…

Cited by 16Source ↗PDF ↗code ↗Scholar ↗
2025

Reliable and Efficient Amortized Model-based Evaluation

Sang T. Truong, Yuheng Tu, Percy Liang, Bo Li, Sanmi Koyejo
ICML 2025poster

Comprehensive evaluations of language models (LM) during both development and deployment phases are necessary because these models are thought to possess numerous capabilities as well as safety risks. The average score across a wide range of benchmarks provides a signal that helps guide the use of t…

Cited by 2Source ↗PDF ↗Scholar ↗
2025

SATBench: Benchmarking LLMs’ Logical Reasoning via Automated Puzzle Generation from SAT Formulas

Anjiang Wei, Yuheng Wu, Yingjia Wan, Tarun Suresh, Huanmi Tan, Zhanke Zhou +3
EMNLP 2025

We introduce SATBench, a benchmark for evaluating the logical reasoning capabilities of large language models (LLMs) through logical puzzles derived from Boolean satisfiability (SAT) problems.Unlike prior work that focuses on inference rule-based reasoning, which often involves deducing conclusions

Cited by 0Source ↗PDF ↗code ↗Scholar ↗
2025

Scaling Laws for Downstream Task Performance in Machine Translation

Berivan Isik, Natalia Ponomareva, Hussein Hazimeh, Dimitris Paparas, Sergei Vassilvitskii, Sanmi Koyejo
ICLR 2025poster

Scaling laws provide important insights that can guide the design of large language models (LLMs). Existing work has primarily focused on studying scaling laws for pretraining (upstream) loss. However, in transfer learning settings, in which LLMs are pretrained on an unsupervised dataset and then fi…

Cited by 2Source ↗PDF ↗Scholar ↗
2025

The Sound of Syntax: Finetuning and Comprehensive Evaluation of Language Models for Speech Pathology

Fagun Patel, Duc Q. Nguyen, Sang T. Truong, Jody Vaynshtok, Sanmi Koyejo, Nick Haber
EMNLP 2025

According to the U.S. National Institutes of Health, more than 3.4 million children experience speech disorders that require clinical intervention. The number of speech-language pathologists (SLPs) is roughly 20 times fewer than the number of affected children, highlighting a significant gap in chil

Cited by 0Source ↗PDF ↗Scholar ↗
2025

The Utility and Complexity of In- and Out-of-Distribution Machine Unlearning

Youssef Allouah, Joshua Kazdan, Rachid Guerraoui, Sanmi Koyejo
ICLR 2025poster

Machine unlearning, the process of selectively removing data from trained models, is increasingly crucial for addressing privacy concerns and knowledge gaps post-deployment. Despite this importance, existing approaches are often heuristic and lack formal guarantees. In this paper, we analyze the fun…

Cited by 1Source ↗PDF ↗Scholar ↗
2025

Understanding challenges to the interpretation of disaggregated evaluations of algorithmic fairness

Stephen R Pfohl, Natalie Harris, Chirag Nagpal, David Madras, Vishwali Mhasawade, Olawale Elijah Salaudeen +9
NeurIPS 2025poster

Disaggregated evaluation across subgroups is critical for assessing the fairness of machine learning models, but its uncritical use can mislead practitioners. We show that equal performance across subgroups is an unreliable measure of fairness when data are representative of the relevant populations…

Cited by 0Source ↗code ↗Scholar ↗
2025

Why Has Predicting Downstream Capabilities of Frontier AI Models with Scale Remained Elusive?

Rylan Schaeffer, Hailey Schoelkopf, Brando Miranda, Gabriel Mukobi, Varun Madan, Adam Ibrahim +3
ICML 2025poster

Predictable behavior from scaling advanced AI systems is an extremely desirable property for engineers, companies, economists and governments alike, and while a well-established literature exists on how pretraining performance scales, predictable scaling behavior on downstream capabilities remains e…

Cited by 14Source ↗PDF ↗Scholar ↗
2024

Adaptive Compression in Federated Learning via Side Information

Berivan Isik, Francesco Pase, Deniz Gunduz, Sanmi Koyejo, Tsachy Weissman, Michele Zorzi
AISTATS 2024poster

The high communication cost of sending model updates from the clients to the server is a significant bottleneck for scalable federated learning (FL). Among existing approaches, state-of-the-art bitrate-accuracy tradeoffs have been achieved using stochastic compression methods – in which the client n…

Cited by 11Source ↗PDF ↗code ↗Scholar ↗
2024

Causally Inspired Regularization Enables Domain General Representations

Olawale Salaudeen, Sanmi Koyejo
AISTATS 2024poster

Given a causal graph representing the data-generating process shared across different domains/distributions, enforcing sufficient graph-implied conditional independencies can identify domain-general (non-spurious) feature representations. For the standard input-output predictive setting, we categori…

Cited by 4Source ↗PDF ↗code ↗Scholar ↗
2024

Crossing Linguistic Horizons: Finetuning and Comprehensive Evaluation of Vietnamese Large Language Models

Sang Truong, Duc Nguyen, Toan Nguyen, Dong Le, Nhi Truong, Tho Quan +1
NAACL 2024findings

Recent advancements in large language models (LLMs) have underscored their importance in the evolution of artificial intelligence. However, despite extensive pretraining on multilingual datasets, available open-sourced LLMs exhibit limited effectiveness in processing Vietnamese. The challenge is exa…

Cited by 5Source ↗PDF ↗code ↗Scholar ↗
2024

Enhancing Robustness of Last Layer Two-Stage Fair Model Corrections

Nathan Stromberg, Rohan Ayyagari, Sanmi Koyejo, Richard Nock, Lalitha Sankar
NeurIPS 2024poster

Last-layer retraining methods have emerged as an efficient framework for correcting existing base models. Within this framework, several methods have been proposed to deal with correcting models for subgroup fairness with and without group membership information. Importantly, prior work has demonstr…

Cited by 1Source ↗PDF ↗Scholar ↗
2024

HIFA: High-fidelity Text-to-3D Generation with Advanced Diffusion Guidance

Junzhe Zhu, Peiye Zhuang, Sanmi Koyejo
ICLR 2024poster

The advancements in automatic text-to-3D generation have been remarkable. Most existing methods use pre-trained text-to-image diffusion models to optimize 3D representations like Neural Radiance Fields (NeRFs) via latent-space denoising score matching. Yet, these methods often result in artifacts an…

Cited by 59Source ↗PDF ↗code ↗Scholar ↗
2024

Implicit Regularization in Feedback Alignment Learning Mechanisms for Neural Networks

Zachary Robertson, Sanmi Koyejo
ICML 2024poster

Feedback Alignment (FA) methods are biologically inspired local learning rules for training neural networks with reduced communication between layers. While FA has potential applications in distributed and privacy-aware ML, limitations in multi-class classification and lack of theoretical understand…

Cited by 0Source ↗PDF ↗Scholar ↗
2024

Invariant Aggregator for Defending against Federated Backdoor Attacks

Xiaoyang Wang, Dimitrios Dimitriadis, Sanmi Koyejo, Shruti Tople
AISTATS 2024poster

Federated learning enables training high-utility models across several clients without directly sharing their private data. As a downside, the federated setting makes the model vulnerable to various adversarial attacks in the presence of malicious clients. Despite the theoretical and empirical succe…

Cited by 5Source ↗PDF ↗Scholar ↗
2024

Principled Federated Domain Adaptation: Gradient Projection and Auto-Weighting

Enyi Jiang, Yibo Jacky Zhang, Sanmi Koyejo
ICLR 2024poster

Federated Domain Adaptation (FDA) describes the federated learning (FL) setting where source clients and a server work collaboratively to improve the performance of a target client where limited data is available. The domain shift between the source and target domains, coupled with limited data of t…

Cited by 9Source ↗PDF ↗code ↗Scholar ↗
2024

Proxy Methods for Domain Adaptation

Katherine Tsai, Stephen R Pfohl, Olawale Salaudeen, Nicole Chiou, Matt Kusner, Alexander D’Amour +2
AISTATS 2024poster

We study the problem of domain adaptation under distribution shift, where the shift is due to a change in the distribution of an unobserved, latent variable that confounds both the covariates and the labels. In this setting, neither the covariate shift nor the label shift assumptions apply. Our appr…

Cited by 5Source ↗PDF ↗code ↗Scholar ↗
2024

Sketching for Distributed Deep Learning: A Sharper Analysis

Mayank Shrivastava, Berivan Isik, Qiaobo Li, Sanmi Koyejo, Arindam Banerjee
NeurIPS 2024poster

The high communication cost between the server and the clients is a significant bottleneck in scaling distributed learning for overparametrized deep models. One popular approach for reducing this communication overhead is randomized sketching. However, existing theoretical analyses for sketching-bas…

Cited by 1Source ↗PDF ↗Scholar ↗
2024

Transforming and Combining Rewards for Aligning Large Language Models

Zihao Wang, Chirag Nagpal, Jonathan Berant, Jacob Eisenstein, Alexander Nicholas D'Amour, Sanmi Koyejo +1
ICML 2024poster

A common approach for aligning language models to human preferences is to first learn a reward model from preference data, and then use this reward model to update the language model. We study two closely related problems that arise in this approach. First, any monotone transformation of the reward…

Cited by 9Source ↗PDF ↗Scholar ↗
2023

Adapting to Latent Subgroup Shifts via Concepts and Proxies

Ibrahim Alabdulmohsin, Nicole Chiou, Alexander D’Amour, Arthur Gretton, Sanmi Koyejo, Matt J. Kusner +4
AISTATS 2023poster

We address the problem of unsupervised domain adaptation when the source domain differs from the target domain because of a shift in the distribution of a latent subgroup. When this subgroup confounds all observed data, neither covariate shift nor label shift assumptions apply. We show that the opti…

Cited by 16Source ↗PDF ↗code ↗Scholar ↗
2023

Cooperative Inverse Decision Theory for Uncertain Preferences

Zachary Robertson, Hantao Zhang, Sanmi Koyejo
AISTATS 2023poster

Inverse decision theory (IDT) aims to learn a performance metric for classification by eliciting expert classifications on examples. However, elicitation in practical settings may require many classifications of potentially ambiguous examples. To improve the efficiency of elicitation, we propose the…

Cited by 0Source ↗PDF ↗code ↗Scholar ↗
2023

DecodingTrust: A Comprehensive Assessment of Trustworthiness in GPT Models

Boxin Wang, Weixin Chen, Hengzhi Pei, Chulin Xie, Mintong Kang, Chenhui Zhang +13
NeurIPS 2023oral

Generative Pre-trained Transformer (GPT) models have exhibited exciting progress in capabilities, capturing the interest of practitioners and the public alike. Yet, while the literature on the trustworthiness of GPT models remains limited, practitioners have proposed employing capable GPT models for…

Cited by 473Source ↗PDF ↗code ↗Scholar ↗
2023

Finite-sample guarantees for Nash Q-learning with linear function approximation

Pedro Cisneros-Velarde, Sanmi Koyejo
UAI 2023poster

Nash Q-learning may be considered one of the first and most known algorithms in multi-agent reinforcement learning (MARL) for learning policies that constitute a Nash equilibrium of an underlying general-sum Markov game. Its original proof provided asymptotic guarantees and was for the tabular case.…

Cited by 3Source ↗PDF ↗Scholar ↗
2023

One Policy is Enough: Parallel Exploration with a Single Policy is Near-Optimal for Reward-Free Reinforcement Learning

Pedro Cisneros-Velarde, Boxiang Lyu, Sanmi Koyejo, Mladen Kolar
AISTATS 2023poster

Although parallelism has been extensively used in Reinforcement Learning (RL), the quantitative effects of parallel exploration are not well understood theoretically. We study the benefits of simple parallel exploration for reward-free RL in linear Markov decision processes (MDPs) and two-player zer…

Cited by 4Source ↗PDF ↗Scholar ↗
2023

Self-Supervised Learning of Representations for Space Generates Multi-Modular Grid Cells

Rylan Schaeffer, Mikail Khona, Tzuhsuan Ma, Cristobal Eyzaguirre, Sanmi Koyejo, Ila R Fiete
NeurIPS 2023poster

To solve the spatial problems of mapping, localization and navigation, the mammalian lineage has developed striking spatial representations. One important spatial representation is the Nobel-prize winning grid cells: neurons that represent self-location, a local and aperiodic quantity, with seemingl…

Cited by 22Source ↗PDF ↗Scholar ↗
2022

Adversarially Robust Models may not Transfer Better: Sufficient Conditions for Domain Transferability from the View of Regularization

Xiaojun Xu, Jacky Y Zhang, Evelyn Ma, Hyun Ho Son, Sanmi Koyejo, Bo Li
ICML 2022spotlight

Machine learning (ML) robustness and domain generalization are fundamentally correlated: they essentially concern data distribution shifts under adversarial and natural settings, respectively. On one hand, recent studies show that more robust (adversarially trained) models are more generalizable. On…

Cited by 13Source ↗PDF ↗Scholar ↗
2021

Bayesian Coresets: Revisiting the Nonconvex Optimization Perspective

Jacky Zhang, Rajiv Khanna, Anastasios Kyrillidis, Sanmi Koyejo
AISTATS 2021poster

Bayesian coresets have emerged as a promising approach for implementing scalable Bayesian inference. The Bayesian coreset problem involves selecting a (weighted) subset of the data samples, such that the posterior inference using the selected subset closely approximates the posterior inference using…

Cited by 27Source ↗PDF ↗code ↗Scholar ↗
2021

Optimizing Black-box Metrics with Iterative Example Weighting

Gaurush Hiranandani, Jatin Mathur, Harikrishna Narasimhan, Mahdi Milani Fard, Sanmi Koyejo
ICML 2021spotlight

We consider learning to optimize a classification metric defined by a black-box function of the confusion matrix. Such black-box learning settings are ubiquitous, for example, when the learner only has query access to the metric of interest, or in noisy-label and domain adaptation applications where…

Cited by 10Source ↗PDF ↗code ↗Scholar ↗
2021

Uncovering the Connections Between Adversarial Transferability and Knowledge Transferability

Kaizhao Liang, Jacky Y Zhang, Boxin Wang, Zhuolin Yang, Sanmi Koyejo, Bo Li
ICML 2021spotlight

Knowledge transferability, or transfer learning, has been widely adopted to allow a pre-trained model in the source domain to be effectively adapted to downstream tasks in the target domain. It is thus important to explore and understand the factors affecting knowledge transferability. In this paper…

Cited by 21Source ↗PDF ↗Scholar ↗
2020

CSER: Communication-efficient SGD with Error Reset

Cong Xie, Shuai Zheng, Sanmi Koyejo, Indranil Gupta, Mu Li, Haibin Lin
NeurIPS 2020poster

The scalability of Distributed Stochastic Gradient Descent (SGD) is today limited by communication bottlenecks. We propose a novel SGD variant: \underline{C}ommunication-efficient \underline{S}GD with \underline{E}rror \underline{R}eset, or \underline{CSER}. The key idea in CSER is first a new techn…

Cited by 50Source ↗PDF ↗Scholar ↗
2020

Fairness with Overlapping Groups; a Probabilistic Perspective

Forest Yang, Mouhamadou Cisse, Sanmi Koyejo
NeurIPS 2020poster

In algorithmically fair prediction problems, a standard goal is to ensure the equality of fairness metrics across multiple overlapping groups simultaneously. We reconsider this standard fair classification problem using a probabilistic population analysis, which, in turn, reveals the Bayes-optimal c…

Cited by 101Source ↗PDF ↗code ↗Scholar ↗
2020

Optimization and Analysis of the pAp@k Metric for Recommender Systems

Gaurush Hiranandani, Warut Vijitbenjaronk, Sanmi Koyejo, Prateek Jain
ICML 2020poster

Modern recommendation and notification systems must be robust to data imbalance, limitations on the number of recommendations/notifications, and heterogeneous engagement profiles across users. The pAp@k metric, which combines the partial-AUC and the precision@k metrics, was recently proposed to eval…

Cited by 9Source ↗PDF ↗code ↗Scholar ↗
2019

Interpreting Black Box Predictions using Fisher Kernels

Rajiv Khanna, Been Kim, Joydeep Ghosh, Sanmi Koyejo
AISTATS 2019poster

Research in both machine learning and psychology suggests that salient examples can help humans to interpret learning models. To this end, we take a novel look at black box interpretation of test predictions in terms of training examples. Our goal is to ask “which training examples are most responsi…

Cited by 123Source ↗PDF ↗Scholar ↗
2019

Max-Sliced Wasserstein Distance and Its Use for GANs

Ishan Deshpande, Yuan-Ting Hu, Ruoyu Sun, Ayis Pyrros, Nasir Siddiqui, Sanmi Koyejo +3
CVPR 2019oral

Generative adversarial nets (GANs) and variational auto-encoders have significantly improved our distribution modeling capabilities, showing promise for dataset augmentation, image-to-image translation and feature learning. However, to model high-dimensional distributions, sequential training and s…

Cited by 238PDF ↗Scholar ↗
2019

Zeno: Distributed Stochastic Gradient Descent with Suspicion-based Fault-tolerance

Cong Xie, Sanmi Koyejo, Indranil Gupta
ICML 2019oral

We present Zeno, a technique to make distributed machine learning, particularly Stochastic Gradient Descent (SGD), tolerant to an arbitrary number of faulty workers. Zeno generalizes previous results that assumed a majority of non-faulty nodes; we need assume only one non-faulty worker. Our key idea…

Cited by 346Source ↗PDF ↗code ↗Scholar ↗
2018

Binary Classification with Karmic, Threshold-Quasi-Concave Metrics

Bowei Yan, Sanmi Koyejo, Kai Zhong, Pradeep Ravikumar
ICML 2018oral

Complex performance measures, beyond the popular measure of accuracy, are increasingly being used in the context of binary classification. These complex performance measures are typically not even decomposable, that is, the loss evaluated on a batch of samples cannot typically be expressed as a sum…

Cited by 36Source ↗PDF ↗Scholar ↗