Knowledge catalogue
arXiv:2601.14104v2 Announce Type: replace-cross Abstract: Backdoor attacks can cause reinforcement learning (RL) policies to behave normally under clean inputs while executing malicious behaviors when
arXiv:2602.22474v2 Announce Type: replace-cross Abstract: Policy steering is an emerging way to adapt robot behaviors at deployment-time: a learned verifier analyzes low-level action samples proposed
arXiv:2605.12780v1 Announce Type: cross Abstract: Calibrated probability outputs of trained classifiers are increasingly used as inputs to downstream regression estimands such as effects, prevalences,
a dude who plays fast and loose with a nonprofit’s money like this is going to play fast and loose with you future. it’s time to boycott OpenAI. 🚨 OpenAI's original board REJECTED Altman's Helion deal
arXiv:2604.00626v2 Announce Type: replace-cross Abstract: As Large Language Models (LLMs) continue to grow in both capability and cost, transferring frontier capabilities into smaller, deployable stud
A survey predicted that 43% of companies would be deploying AI at scale by the current date, but actual adoption has reached only 19%, indicating a significant gap between forecasted and realized AI i
arXiv:2605.11302v1 Announce Type: cross Abstract: We study language generation in the limit under a global preference ordering on strings, as introduced by Kleinberg and Wei. As in [arXiv:2504.14370,
arXiv:2605.12197v1 Announce Type: new Abstract: Leveraging Graph Neural Networks (GNNs) as graph encoders and aligning the resulting representations with Large Language Models (LLMs) through alignment
arXiv:2605.11009v1 Announce Type: new Abstract: Long-horizon, sparse-reward tasks pose a fundamental challenge for reinforcement learning, since single-step TD learning suffers from bootstrapping erro
Russell Brandom / TechCrunch: Adaption, co-founded by ex-Cohere VP of AI research Sara Hooker, unveils AutoScientist, which can automate the research loop behind model training and alignment — For yea
arXiv:2605.11191v1 Announce Type: cross Abstract: Adaptive experimentation under unknown network interference requires solving two coupled problems: (i) learning the underlying dynamics of interferenc
arXiv:2605.11880v1 Announce Type: new Abstract: TD(lambda) in value-based MARL algorithms or the Temporal Difference critic learning in Actor-Critic-based (AC-based) algorithms synergistically integra
arXiv:2605.11458v1 Announce Type: cross Abstract: On-policy self-distillation has become a strong recipe for LLM reasoning, where a privileged teacher supervises the student's own rollouts while condi
agreed. RL is not (at least by itself) the way to alignment Yoshua Bengio says Reinforcement Learning is a dangerous path for building superintelligence It can create systems with hidden goals, reward
arXiv:2511.22663v5 Announce Type: replace Abstract: Unified multimodal models for image generation and understanding represent a significant step toward AGI and have attracted widespread attention fro
.@alexsobel's new AI kill-switch amendment gives government the power to shut down data centers in cases of AI emergency. It's also the first piece of proposed UK law recognizing ASI as the national s
arXiv:2605.12416v1 Announce Type: new Abstract: Generative policies based on expressive model classes, such as diffusion and flow matching, are well-suited to complex control problems with highly mult
arXiv:2605.12495v1 Announce Type: new Abstract: In this paper, we propose AlphaGRPO, a novel framework that applies Group Relative Policy Optimization (GRPO) to AR-Diffusion Unified Multimodal Models
arXiv:2605.11609v1 Announce Type: cross Abstract: On-policy self-distillation, where a student is pulled toward a copy of itself conditioned on privileged context (e.g., a verified solution or feedbac
arXiv:2605.11968v1 Announce Type: cross Abstract: Modern Earth System Models (ESMs) operate on horizontal scales far larger than typical cloud features, requiring stochastic subcolumn generators to re
arXiv:2604.04894v2 Announce Type: replace Abstract: Reinforcement learning with verifiable rewards (RLVR) has substantially improved the reasoning ability of large language models (LLMs), but it often
arXiv:2605.11694v1 Announce Type: new Abstract: We study policy optimization for infinite-horizon, discounted constrained Markov decision processes (CMDPs). While existing theoretical guarantees typic
arXiv:2605.12316v1 Announce Type: new Abstract: We study the fundamental and timely problem of learning long sequences in autoregressive modeling and next-token prediction under model misspecification
arXiv:2605.11904v1 Announce Type: new Abstract: The Nearest Class Mean (NCM) classifier is widely favored in Class-Incremental Learning (CIL) for its superior resistance to catastrophic forgetting com
arXiv:2605.11300v1 Announce Type: new Abstract: Vision state space models inherit the efficiency and long-range modeling ability of Mamba-style selective scans. However, their performance depends crit
arXiv:2605.11365v1 Announce Type: cross Abstract: Automated systems built on artificial intelligence (AI) are increasingly deployed across high-stakes domains, raising critical concerns about fairness
arXiv:2605.11362v1 Announce Type: new Abstract: In the data-driven era, large-scale datasets are routinely collected and analyzed using machine learning (ML) and artificial intelligence (AI) to inform
arXiv:2605.11304v1 Announce Type: new Abstract: Chest radiograph interpretation requires temporal reasoning over prior and current studies, yet most vision-language models are trained on static image-
arXiv:2601.21944v2 Announce Type: replace Abstract: The widespread adoption of deep learning models in computer vision has intensified concerns about interpretability. Despite strong performance, thes
arXiv:2605.11939v1 Announce Type: new Abstract: Prompt learning has emerged as an efficient alternative to fine-tuning pre-trained vision-language models (VLMs). Despite its promise, current methods s
arXiv:2605.12227v1 Announce Type: new Abstract: Adapting large language models (LLMs) to long-context tasks requires post-training methods that remain accurate and coherent over thousands of tokens. E
arXiv:2605.11519v1 Announce Type: cross Abstract: Using offline datasets to evaluate conversational agents often fails to cover rare scenarios or to support testing new policies. This has motivated th
arXiv:2605.11485v1 Announce Type: new Abstract: Imitation learning powered by generative models has proven effective for modeling complex single-agent behaviors. However, teaching multi-agent systems,
counterpoint: i have no interest in a company that says we will roll everything out to anybody regardless of how dangerous it is. boris cherny recently says that, 'anthropic has no plans to roll out m
arXiv:2605.11311v1 Announce Type: cross Abstract: Diffusion models typically generate image batches from independent Gaussian initial noises. We argue that this independence assumption is only one cho
arXiv:2605.12145v1 Announce Type: new Abstract: Multimodal learning seeks to integrate information across diverse sensory sources, yet current approaches struggle to balance cross-modal generalizabili
arXiv:2605.11711v1 Announce Type: new Abstract: Model-based representations recently stand out as a promising framework that embeds latent dynamics information into the representations for downstream
arXiv:2605.11145v1 Announce Type: cross Abstract: Collaborative filtering (CF) models based on graph neural networks (GNNs) achieve strong performance in recommender systems by propagating user-item s
arXiv:2605.11764v1 Announce Type: new Abstract: Machine-learning predictors of biochemical activity often exhibit large random-split-to-leave-one-target-out generalisation gaps that have been document
arXiv:2602.06462v3 Announce Type: replace Abstract: Masked diffusion language models generate text through iterative masked-token filling, but terminal-only rewards on final completions provide coarse
arXiv:2605.12379v1 Announce Type: new Abstract: Many reinforcement learning (RL) tasks have discrete action spaces, but most generative policy methods based on diffusion and flow matching are designed
arXiv:2509.09838v2 Announce Type: replace Abstract: While Soft Actor-Critic (SAC) is highly effective in continuous control, its discrete counterpart (DSAC) performs poorly on challenging discrete-act
Do multi-agent systems make LLM reasoning better? Most AI devs assume that it should. But this new paper shows that this is often not the case. It ran 22,500 deterministic trajectories across GAIA, SW
arXiv:2505.18780v3 Announce Type: replace-cross Abstract: Achieving versatile humanoid locomotion with a single policy presents a critical scalability challenge. Prevailing methods often rely on disti
arXiv:2605.12196v1 Announce Type: new Abstract: Accurate ultra-short-term wind power forecasting is critical for grid dispatch and reserve management, yet remains challenging due to the non-stationary
arXiv:2605.12335v1 Announce Type: cross Abstract: Electronic Health Records (EHR) contain rich longitudinal patient information and are widely used in predictive modeling applications. However, effect
arXiv:2605.11695v1 Announce Type: new Abstract: Symbols are shared, but perception is private. We study emergent communication between heterogeneous visual agents through decentralized learning, askin
arXiv:2602.22347v2 Announce Type: replace Abstract: Foundation models for computational pathology are expected to facilitate the development of high-performing, generalisable deep learning systems. Ho
arXiv:2605.11093v1 Announce Type: new Abstract: Today's inference-time workloads increasingly depend on timely access to a model's internal states. We present DMI-Lib, a high-speed deep model inspecto
arXiv:2605.11214v1 Announce Type: new Abstract: Hard constraints in generative sampling are typically enforced by projection, applied either once at the end of sampling or after every update. This bin
arXiv:2605.11632v1 Announce Type: new Abstract: Self-generated counterfactual explanations (SCEs) are minimally modified inputs (minimality) generated by large language models (LLMs) that flip their o
arXiv:2605.11964v1 Announce Type: new Abstract: A target-guided proactive dialogue system aims to steer conversations proactively toward pre-defined targets, such as designated keywords or specific to
arXiv:2605.11775v1 Announce Type: cross Abstract: Policy entropy has emerged as a fundamental measure for understanding and controlling exploration in reinforcement learning with verifiable rewards (R
arXiv:2605.11328v1 Announce Type: new Abstract: Automated scientific discovery using large language models relies on identifying genuinely novel solutions. Standard reinforcement learning penalizes hi
arXiv:2605.12388v1 Announce Type: cross Abstract: Effective multi-agent cooperation requires agents to adopt diverse behaviors as task conditions evolve-and to do so at the right moment. Yet, current
arXiv:2605.11859v1 Announce Type: new Abstract: Robot navigation is a crucial task with applications to social robots in dynamic human environments. While Reinforcement Learning (RL) has shown great p
arXiv:2605.12174v1 Announce Type: new Abstract: Solving optimal transport (OT) on random minibatches is a common surrogate for exact OT in large-scale learning. In flow matching (FM), this surrogate i
arXiv:2605.12195v1 Announce Type: new Abstract: Conformal prediction methods provide statistically rigorous marginal coverage guarantees for machine learning models, but such guarantees fail to accoun
arXiv:2605.11815v1 Announce Type: new Abstract: Hierarchical federated learning (HFL) leverages edge servers for partial aggregation in edge computing. Yet existing FL methods lack mechanisms for join
arXiv:2605.11571v1 Announce Type: new Abstract: Federated learning usually aggregates client updates using dataset size or gradient-level criteria, while overlooking internal signals about how each cl