Congratulations to the #ICML2026 award winners

This year’s International Conference on Machine Learning (ICML2025) took place in Seoul, South Korea, from 6-11 July. The winners of the 2026 outstanding paper, outstanding position paper, and test-of-time awards were announced during the event.

Outstanding paper award

There are two outstanding papers this year:

The Flexibility Trap: Rethinking the Value of Arbitrary Order in Diffusion Language Models
Zanlin Ni, Shenzhi Wang, Yang Yue, Tianyu Yu, Weilin Zhao, Yeguo Hua, Tianyi Chen, Jun Song, YuCheng, Bo Zheng, Gao Huang

Abstract: Diffusion Large Language Models (dLLMs) break the rigid left-to-right constraint of traditional LLMs, enabling token generation in arbitrary orders. Intuitively, this flexibility implies a solution space that strictly supersets the fixed autoregressive trajectory, theoretically unlocking superior reasoning potential. Indeed, for specific constraint satisfaction tasks (e.g., sudoku puzzles), this capability has proven to be highly advantageous. However, in this paper, we reveal that for general reasoning tasks (e.g., mathematics and coding), arbitrary order generation may in fact limit the reasoning potential of dLLMs. We find that dLLMs tend to exploit this order flexibility to bypass high-uncertainty tokens that are crucial for exploration, leading to a premature collapse of solution coverage. This observation motivates a rethink of RL approaches for dLLMs, where considerable complexities, such as handling combinatorial trajectories and intractable likelihoods, are often devoted to preserving this flexibility. We demonstrate that effective reasoning can be better elicited by simply forgoing arbitrary order and applying standard Group Relative Policy Optimization (GRPO) instead. Our approach, JustGRPO, is minimalist yet surprisingly effective (e.g., 89.1% accuracy on GSM8K) while fully retaining the parallel decoding ability of dLLMs.

Read the paper in full here.

High-accuracy sampling for diffusion models and log-concave distributions
Fan Chen, Sinho Chewi, Constantinos Daskalakis, Alexander Rakhlin

Abstract: We present algorithms for diffusion model sampling which obtain δ-error in polylog(1/δ) steps, given access to O˜(δ)-accurate score estimates in L2. This is an exponential improvement over all previous results. Specifically, under minimal data assumptions, the complexity is O˜(d⋆polylog(1/δ)) where d⋆ is the intrinsic dimension of the data. Further, under a non-uniform L-Lipschitz condition, the complexity reduces to O˜(Lpolylog(1/δ)). Our approach also yields the first polylog(1/δ) complexity sampler for general log-concave distributions using only gradient evaluations.

Read the paper in full here.

Outstanding paper award honourable mentions

Five papers received honourable mentions this year:

To Grok Grokking: Provable Grokking in Ridge Regression
Mingyue Xu, Gal Vardi, Itay Safran

Abstract: We study grokking – the onset of generalization long after overfitting – in a classical ridge regression setting. We prove end-to-end grokking results for learning over-parameterized linear regression models using gradient descent with weight decay. Specifically, we prove that the following stages occur: (i) the model overfits the training data early during training; (ii) poor generalization persists long after overfitting has manifested; and (iii) the generalization error eventually becomes arbitrarily small. Moreover, we show, both theoretically and empirically, that grokking can be amplified or eliminated in a principled manner through proper hyperparameter tuning. To the best of our knowledge, these are the first rigorous quantitative bounds on the generalization delay (which we refer to as the “grokking time”) in terms of training hyperparameters. Lastly, going beyond the linear setting, we empirically demonstrate that our quantitative bounds also capture the behavior of grokking on non-linear neural networks. Our results suggest that grokking is not an inherent failure mode of deep learning, but rather a consequence of specific training conditions, and thus does not require fundamental changes to the model architecture or learning algorithm to avoid.

Read the paper in full here.

The Obfuscation Atlas: Mapping Where Honesty Emerges in RLVR with Deception Probes
Mohammad Taufeeque, Stefan Heimersheim, Adam Gleave, Chris Cundy

Abstract: Training against white-box deception detectors has been proposed as a way to make AI systems honest. However, such training risks models learning to obfuscate their deception to evade the detector. Prior work has studied obfuscation only in artificial settings where models were directly rewarded for harmful output. We construct a realistic coding environment where reward hacking via hardcoding test cases naturally occurs, and show that obfuscation emerges in this setting. We introduce a taxonomy of possible outcomes when training against a deception detector. The model either remains honest, or becomes deceptive via two possible obfuscation strategies. (i) Obfuscated activations: the model outputs deceptive text while its activations change to no longer trigger the detector. (ii) Obfuscated policy: the model produces detector-evading deceptive text, typically by including a justification for the reward hack. Empirically, obfuscated activations arise from representation drift during RL, with or without a detector penalty. The penalty only incentivizes obfuscated policies: we theoretically show this is expected for policy gradient methods. Sufficiently high KL regularization and detector penalty reliably yield honest policies, establishing white-box deception detectors as viable training signals for tasks prone to reward hacking.

Read the paper in full here.

How much can language models memorize?
John X. Morris, Chawin Sitawarin, Chuan Guo, Narine Kokhlikyan, G. Edward Suh, Alexander M. Rush, Kamalika Chaudhuri, Saeed Mahloujifar

Abstract: We propose a new method for estimating how much a model knows about a datapoint and use it to measure the capacity of modern language models. Prior studies of language model memorization have struggled to disentangle memorization from generalization. We formally separate memorization into two components: unintended memorization, the information a model contains about a specific dataset, and generalization, the information a model contains about the true data-generation process. When we completely eliminate generalization, we can compute the total memorization, which provides an estimate of model capacity: our measurements estimate that GPT-style models have a capacity of approximately 3.6 bits per parameter. We train language models on datasets of increasing size and observe that models memorize until their capacity fills, at which point unintended memorization decreases as models begin to generalize. We train hundreds of transformer language models ranging from 500K to 1.5B parameters and produce a series of scaling laws relating model capacity and data size to membership inference.

Read the paper in full here.

Motion Attribution for Video Generation
Xindi Wu, Despoina Paschalidou, Jun Gao, Antonio Torralba, Laura Leal-Taixé, Olga Russakovsky, Sanja Fidler, Jonathan Lorraine

Abstract: Despite the rapid progress of video generation models, the role of data in influencing motion is poorly understood. We present Motive (MOTIon attribution for Video gEneration), a motion-centric, gradient-based data attribution framework that scales to modern, large, high-quality video datasets and models. We use this to study which fine-tuning clips improve or degrade temporal dynamics. Motive isolates temporal dynamics from static appearance via motion-weighted loss masks, yielding efficient and scalable motion-specific influence computation. On text-to-video models, Motive identifies clips that strongly affect motion and guides data curation that improves temporal consistency and physical plausibility. With Motive-selected high-influence data, we improve both motion smoothness and dynamic degree on VBench, achieving a 74.1% human preference win rate compared with the pretrained base model. To our knowledge, this is the first framework to attribute motion rather than visual appearance in video generative models and to use it to curate fine-tuning data.

Read the paper in full here.

A Random Matrix Perspective on the Consistency of Diffusion Models
Binxu Wang, Jacob Zavatone-Veth, Cengiz Pehlevan

Abstract: Diffusion models trained on different, non-overlapping subsets of a dataset often produce strikingly similar outputs when given the same noise seed. We trace this consistency to a simple linear effect: the shared Gaussian statistics across splits already predict much of the generated images. To formalize this, we develop a random matrix theory (RMT) framework that quantifies how finite datasets shape the expectation and variance of the learned denoiser and sampling map in the linear setting. For expectations, sampling variability acts as a renormalization of the noise level through a self-consistent relation σ2↦κ(σ2), explaining why limited data overshrink low-variance directions and pull samples toward the dataset mean. For fluctuations, our variance formulas reveal three key factors behind cross-split disagreement: \textit{anisotropy} across eigenmodes, \textit{inhomogeneity} across inputs, and overall scaling with dataset size. Extending deterministic-equivalence tools to fractional matrix powers further allows us to analyze entire sampling trajectories. The theory sharply predicts the behavior of linear diffusion models, and we validate its predictions on UNet and DiT architectures in their non-memorization regime, identifying where and how samples deviates across training data split. This provides a principled baseline for reproducibility in diffusion training, linking spectral properties of data to the stability of generative outputs.

Read the paper in full here.

Outstanding position paper award

The winner of the outstanding position paper award is:

Position: The Alignment Community is Unintentionally Building a Censor’s Toolkit
Sarah Ball, Phil Hackemann

Abstract: This position paper argues that modern alignment methods – originally designed to prevent harmful output – are dual-use technologies that may easily be misused by malicious actors for censorship and manipulation. By mapping current alignment techniques to the possibility and actual cases of misuse, we show that the quest for a ”perfectly aligned” model inadvertently also provides malicious actors with an ever-improving tool for informational dominance. We need to discuss this dual-use potential now, as its risk is exacerbated by rapid user adoption of AI as information provider and a political landscape that increasingly shifts towards authoritarianism. We conclude by urging the community to consider the intentional misuse of safety mechanisms and propose mitigation strategies to safeguard against this dual-use potential.

Read the paper in full here.

Outstanding position paper award honourable mention

The follow paper received an honourable mention in the outstanding position paper category:

Position: AI/ML Deepfake Research is Misaligned with AI Generated Non-Consensual Intimate Imagery (AIG-NCII)
Li Qiwei, Wells Lucas Santo, Sarita Schoenebeck, Eric Gilbert

Abstract: AI-generated non-consensual intimate imagery (AIG-NCII) is not adequately addressed in AI/ML literature regarding AI-generated media, commonly referred to as “deepfakes”. While research on deepfakes currently focuses on its epistemic harms—or harms relating to truth and authenticity—this is misaligned with the dominant reality of generative AI abuse involving sexualized imagery. We conduct a landscape analysis of highly-cited works to demonstrate that technical interventions addressing deepfakes almost entirely ignore AIG-NCII, limiting the research ecosystem to authenticity detection tools. In this position paper, we argue that existing interventions address viewer-centric epistemic harms, such as fraud or scams, but ignore subject-centric dignity harms, such as AIG-NCII. We illustrate that knowing an image is synthetic does not mitigate harms to subjects and may, in some cases, even exacerbate them. We conclude by offering recommendations to realign the field, including updating threat models to consider subject-centered harms and addressing AIG-NCII in AI safety research. Finally, we caution that researchers should only engage in this high-risk domain if they implement safety guardrails for both subjects and researchers and establish partnerships with domain experts in sexual violence prevention.

Read the paper in full here.

Test-of-time award

The test-of-time award celebrates the lasting impact a paper has had over the past ten years. For this year’s award paper that were presented at ICML 2016 have been considered. There is one winner this year:

Asynchronous Methods for Deep Reinforcement Learning
Volodymyr Mnih, Adrià Puigdomènech Badia, Mehdi Mirza, Alex Graves, Timothy P. Lillicrap, Tim Harley, David Silver, Koray Kavukcuoglu

Abstract: We propose a conceptually simple and lightweight framework for deep reinforcement learning that uses asynchronous gradient descent for optimization of deep neural network controllers. We present asynchronous variants of four standard reinforcement learning algorithms and show that parallel actor-learners have a stabilizing effect on training allowing all four methods to successfully train neural network controllers. The best performing method, an asynchronous variant of actor-critic, surpasses the current state-of-the-art on the Atari domain while training for half the time on a single multi-core CPU instead of a GPU. Furthermore, we show that asynchronous actor-critic succeeds on a wide variety of continuous motor control problems as well as on a new task of navigating random 3D mazes using a visual input.

Read the paper in full here.

Lucy Smith is Senior Managing Editor for AIhub.

Lucy Smith is Senior Managing Editor for AIhub.