Week beginning 2nd February 2026

Multimodal large language models (MLLMs) have rapidly advanced, yet their adoption in medicine remains limited by gaps in domain coverage, modality alignment, and grounded reasoning. In this work, we introduce MedMO, a medical foundation model built upon a generalized MLLM architecture and trained exclusively on large-scale, domain-specific data. MedMO follows a multi-stage training recipe: (i) cross-modal pretraining to align heterogeneous visual encoders with a medical language backbone; (ii) instruction tuning on multi-task supervision that spans captioning, VQA, report generation, retrieval, and grounded disease localization with bounding boxes; and (iii) reinforcement learning with verifiable rewards that combine factuality checks with a box-level GIoU reward to strengthen spatial grounding and step-by-step reasoning in complex clinical scenarios. MedMO consistently outperforms strong open-source medical MLLMs across multiple modalities and tasks. On VQA benchmarks, MedMO achieves an average accuracy improvement of +13.7% over the baseline and performs within 1.9% of the SOTA Fleming-VL. For text-based QA, it attains +6.9% over the baseline and +14.5% over Fleming-VL. In medical report generation, MedMO delivers significant gains in both semantic and clinical accuracy. Moreover, it exhibits strong grounding capability, achieving an IoU improvement of +40.4 over the baseline and +37.0% over Fleming-VL, underscoring its robust spatial reasoning and localization performance. Evaluations across radiology, ophthalmology, and pathology-microscopy confirm MedMO's broad cross-modality generalization. We release two versions of MedMO: 4B and 8B. Project is available at https://genmilab.github.io/MedMO-Page

Existing approaches for analyzing neural network activations, such as PCA and sparse autoencoders, rely on strong structural assumptions. Generative models offer an alternative: they can uncover structure without such assumptions and act as priors that improve intervention fidelity. We explore this direction by training diffusion models on one billion residual stream activations, creating "meta-models" that learn the distribution of a network's internal states. We find that diffusion loss decreases smoothly with compute and reliably predicts downstream utility. In particular, applying the meta-model's learned prior to steering interventions improves fluency, with larger gains as loss decreases. Moreover, the meta-model's neurons increasingly isolate concepts into individual units, with sparse probing scores that scale as loss decreases. These results suggest generative meta-models offer a scalable path toward interpretability without restrictive structural assumptions. Project page: https://generative-latent-prior.github.io.

Large reasoning models achieve strong performance by scaling inference-time chain-of-thought, but this paradigm suffers from quadratic cost, context length limits, and degraded reasoning due to lost-in-the-middle effects. Iterative reasoning mitigates these issues by periodically summarizing intermediate thoughts, yet existing methods rely on supervised learning or fixed heuristics and fail to optimize when to summarize, what to preserve, and how to resume reasoning. We propose InftyThink+, an end-to-end reinforcement learning framework that optimizes the entire iterative reasoning trajectory, building on model-controlled iteration boundaries and explicit summarization. InftyThink+ adopts a two-stage training scheme with supervised cold-start followed by trajectory-level reinforcement learning, enabling the model to learn strategic summarization and continuation decisions. Experiments on DeepSeek-R1-Distill-Qwen-1.5B show that InftyThink+ improves accuracy by 21% on AIME24 and outperforms conventional long chain-of-thought reinforcement learning by a clear margin, while also generalizing better to out-of-distribution benchmarks. Moreover, InftyThink+ significantly reduces inference latency and accelerates reinforcement learning training, demonstrating improved reasoning efficiency alongside stronger performance.

Cinematic video production requires control over scene-subject composition and camera movement, but live-action shooting remains costly due to the need for constructing physical sets. To address this, we introduce the task of cinematic video generation with decoupled scene context: given multiple images of a static environment, the goal is to synthesize high-quality videos featuring dynamic subject while preserving the underlying scene consistency and following a user-specified camera trajectory. We present CineScene, a framework that leverages implicit 3D-aware scene representation for cinematic video generation. Our key innovation is a novel context conditioning mechanism that injects 3D-aware features in an implicit way: By encoding scene images into visual representations through VGGT, CineScene injects spatial priors into a pretrained text-to-video generation model by additional context concatenation, enabling camera-controlled video synthesis with consistent scenes and dynamic subjects. To further enhance the model's robustness, we introduce a simple yet effective random-shuffling strategy for the input scene images during training. To address the lack of training data, we construct a scene-decoupled dataset with Unreal Engine 5, containing paired videos of scenes with and without dynamic subjects, panoramic images representing the underlying static scene, along with their camera trajectories. Experiments show that CineScene achieves state-of-the-art performance in scene-consistent cinematic video generation, handling large camera movements and demonstrating generalization across diverse environments.

We prove a strongly polynomial bound on the circuit diameter of polyhedra, resolving the circuit analogue of the polynomial Hirsch conjecture. Specifically, we show that the circuit diameter of a polyhedron $P = \{x\in \mathbb{R}^n:\, A x = b, \, x \ge 0\}$ with $A\in\mathbb{R}^{m\times n}$ is $O(m^2 \log m)$. Our construction yields monotone circuit walks, giving the same bound for the monotone circuit diameter. The circuit diameter, introduced by Borgwardt, Finhold, and Hemmecke (SIDMA 2015), is a natural relaxation of the combinatorial diameter that allows steps along circuit directions rather than only along edges. All prior upper bounds on the circuit diameter were only weakly polynomial. Finding a circuit augmentation algorithm that matches this bound would yield a strongly polynomial time algorithm for linear programming, resolving Smale's 9th problem.

Addressing class imbalance is a central challenge in credit card fraud detection, as it directly impacts predictive reliability in real-world financial systems. To overcome this, the study proposes an enhanced workflow based on the Explainable Boosting Machine (EBM)-a transparent, state-of-the-art implementation of the GA2M algorithm-optimized through systematic hyperparameter tuning, feature selection, and preprocessing refinement. Rather than relying on conventional sampling techniques that may introduce bias or cause information loss, the optimized EBM achieves an effective balance between accuracy and interpretability, enabling precise detection of fraudulent transactions while providing actionable insights into feature importance and interaction effects. Furthermore, the Taguchi method is employed to optimize both the sequence of data scalers and model hyperparameters, ensuring robust, reproducible, and systematically validated performance improvements. Experimental evaluation on benchmark credit card data yields an ROC-AUC of 0.983, surpassing prior EBM baselines (0.975) and outperforming Logistic Regression, Random Forest, XGBoost, and Decision Tree models. These results highlight the potential of interpretable machine learning and data-driven optimization for advancing trustworthy fraud analytics in financial systems.

Diffusion large language models (dLLMs) have shown advantages in text generation, particularly due to their inherent ability for parallel decoding. However, constrained by the quality--speed trade-off, existing inference solutions adopt conservative parallel strategies, leaving substantial efficiency potential underexplored. A core challenge is that parallel decoding assumes each position can be filled independently, but tokens are often semantically coupled. Thus, the correct choice at one position constrains valid choices at others. Without modeling these inter-token dependencies, parallel strategies produce deteriorated outputs. Motivated by this insight, we propose DAWN, a training-free, dependency-aware decoding method for fast dLLM inference. DAWN extracts token dependencies and leverages two key motivations: (1) positions dependent on unmasked certain positions become more reliable, (2) simultaneously unmasking strongly coupled uncertain positions induces errors. Given those findings, DAWN leverages a dependency graph to select more reliable unmasking positions at each iteration, achieving high parallelism with negligible loss in generation quality. Extensive experiments across multiple models and datasets demonstrate that DAWN speedups the inference by 1.80-8.06x over baselines while preserving the generation quality. Code is released at https://github.com/lizhuo-luo/DAWN.

Say you and some friends decide to make brackets for March Madness and are told how each of your brackets scored. The question we ask is: when can you determine how the actual tournament went given your scores? We determine the exact minimum number of brackets needed to do this for any March Madness-style tournament regardless of the scoring system used, and more generally we prove effective bounds for the problem for arbitrary single-elimination tournaments.

Being able to simulate the outcomes of actions in varied environments will revolutionize the development of generalist agents at scale. However, modeling these world dynamics, especially for dexterous robotics tasks, poses significant challenges due to limited data coverage and scarce action labels. As an endeavor towards this end, we introduce DreamDojo, a foundation world model that learns diverse interactions and dexterous controls from 44k hours of egocentric human videos. Our data mixture represents the largest video dataset to date for world model pretraining, spanning a wide range of daily scenarios with diverse objects and skills. To address the scarcity of action labels, we introduce continuous latent actions as unified proxy actions, enhancing interaction knowledge transfer from unlabeled videos. After post-training on small-scale target robot data, DreamDojo demonstrates a strong understanding of physics and precise action controllability. We also devise a distillation pipeline that accelerates DreamDojo to a real-time speed of 10.81 FPS and further improves context consistency. Our work enables several important applications based on generative world models, including live teleoperation, policy evaluation, and model-based planning. Systematic evaluation on multiple challenging out-of-distribution (OOD) benchmarks verifies the significance of our method for simulating open-world, contact-rich tasks, paving the way for general-purpose robot world models.

Can AI agents predict whether they will succeed at a task? We study agentic uncertainty by eliciting success probability estimates before, during, and after task execution. All results exhibit agentic overconfidence: some agents that succeed only 22% of the time predict 77% success. Counterintuitively, pre-execution assessment with strictly less information tends to yield better discrimination than standard post-execution review, though differences are not always significant. Adversarial prompting reframing assessment as bug-finding achieves the best calibration.

The usual semantics of multi-agent epistemic logic is based on Kripke models, defined in terms of binary relations on a set of possible worlds. Recently, there has been a growing interest in using simplicial complexes rather than graphs, as models for multi-agent epistemic logic. This approach uses agents' views as the fundamental object instead of worlds. A set of views by different agents about a world forms a simplex, and a set of simplexes defines a simplicial complex, that can serve as a model for multi-agent epistemic logic. This new approach reveals topological information that is implicit in Kripke models, because the binary indistinguishability relations are more clearly seen as n-ary relations in the simplicial complex. This paper, written for an economics audience, introduces simplicial models to non-experts and connects distributed computing, epistemic logic and topology. Our focus is on distributed knowledge and its fixed point, common distributed knowledge. These concepts arise when considering the knowledge that a group of agents would acquire, if they could communicate their local knowledge perfectly. While common knowledge has been shown to be related to consensus, we illustrate how distributed knowledge is related to a task weaker to consensus, called majority consensus. We describe three models of communication, some well-known (immediate snapshot), and others less studied (related to broadcast and test-and-set). When majority consensus is solvable, we describe the distributed knowledge that is used to solve it. When it is not solvable, we present a logical obstruction, a formula that should always be known according to the task specification, but which the players cannot know.

This paper presents the design and implementation of data-driven optimal derivative feedback controllers for an active magnetic levitation system. A direct, model-free control design method based on the reinforcement learning framework is compared with an indirect optimal control design derived from a numerically identified mathematical model of the system. For the direct model-free approach, a policy iteration procedure is proposed, which adds an iteration layer called the epoch loop to gather multiple sets of process data, providing a more diverse dataset and helping reduce learning biases. This direct control design method is evaluated against a comparable optimal control solution designed from a plant model obtained through the combined Dynamic Mode Decomposition with Control (DMDc) and Prediction Error Minimization (PEM) system identification. Results show that while both controllers can stabilize and improve the performance of the magnetic levitation system when compared to controllers designed from a nominal model, the direct model-free approach consistently outperforms the indirect solution when multiple epochs are allowed. The iterative refinement of the optimal control law over the epoch loop provides the direct approach a clear advantage over the indirect method, which relies on a single set of system data to determine the identified model and control.

Tokenization is a pivotal design choice for neural language modeling in morphologically rich languages (MRLs) such as Turkish, where productive agglutination challenges both vocabulary efficiency and morphological fidelity. Prior studies have explored tokenizer families and vocabulary sizes but typically (i) vary vocabulary without systematically controlling the tokenizer's training corpus, (ii) provide limited intrinsic diagnostics, and (iii) evaluate a narrow slice of downstream tasks. We present the first comprehensive, principled study of Turkish subword tokenization; a "subwords manifest", that jointly varies vocabulary size and tokenizer training corpus size (data and vocabulary coupling), compares multiple tokenizer families under matched parameter budgets (WordPiece, morphology level, and character baselines), and evaluates across semantic (NLI, STS, sentiment analysis, NER), syntactic (POS, dependency parsing), and morphology-sensitive probes. To explain why tokenizers succeed or fail, we introduce a morphology-aware diagnostic toolkit that goes beyond coarse aggregates to boundary-level micro/macro F1, decoupled lemma atomicity vs. surface boundary hits, over/under-segmentation indices, character/word edit distances (CER/WER), continuation rates, and affix-type coverage and token-level atomicity. Our contributions are fourfold: (i) a systematic investigation of the vocabulary-corpus-success triad; (ii) a unified, morphology-aware evaluation framework linking intrinsic diagnostics to extrinsic outcomes; (iii) controlled comparisons identifying when character-level and morphology-level tokenization pay off; and (iv) an open-source release of evaluation code, tokenizer pipelines, and models. As the first work of its kind, this "subwords manifest" delivers actionable guidance for building effective tokenizers in MRLs and establishes a reproducible foundation for future research.

Large language models can resist task-misaligned activation steering during inference, sometimes recovering mid-generation to produce improved responses even when steering remains active. We term this Endogenous Steering Resistance (ESR). Using sparse autoencoder (SAE) latents to steer model activations, we find that Llama-3.3-70B shows substantial ESR, while smaller models from the Llama-3 and Gemma-2 families exhibit the phenomenon less frequently. We identify 26 SAE latents that activate differentially during off-topic content and are causally linked to ESR in Llama-3.3-70B. Zero-ablating these latents reduces the multi-attempt rate by 25%, providing causal evidence for dedicated internal consistency-checking circuits. We demonstrate that ESR can be deliberately enhanced through both prompting and training: meta-prompts instructing the model to self-monitor increase the multi-attempt rate by 4x for Llama-3.3-70B, and fine-tuning on self-correction examples successfully induces ESR-like behavior in smaller models. These findings have dual implications: ESR could protect against adversarial manipulation but might also interfere with beneficial safety interventions that rely on activation steering. Understanding and controlling these resistance mechanisms is important for developing transparent and controllable AI systems. Code is available at github.com/agencyenterprise/endogenous-steering-resistance.

Learning unsupervised representations that are both semantically meaningful and stable across runs remains a central challenge in modern representation learning. We introduce entropy-ordered flows (EOFlows), a normalizing-flow framework that orders latent dimensions by their explained entropy, analogously to PCA's explained variance. This ordering enables adaptive injective flows: after training, one may retain only the top C latent variables to form a compact core representation while the remaining variables capture fine-grained detail and noise, with C chosen flexibly at inference time rather than fixed during training. EOFlows build on insights from Independent Mechanism Analysis, Principal Component Flows and Manifold Entropic Metrics. We combine likelihood-based training with local Jacobian regularization and noise augmentation into a method that scales well to high-dimensional data such as images. Experiments on the CelebA dataset show that our method uncovers a rich set of semantically interpretable features, allowing for high compression and strong denoising.

Non-Markovian dynamics are commonly found in real-world environments due to long-range dependencies, partial observability, and memory effects. The Bellman equation that is the central pillar of Reinforcement learning (RL) becomes only approximately valid under Non-Markovian. Existing work often focus on practical algorithm designs and offer limited theoretical treatment to address key questions, such as what dynamics are indeed capturable by the Bellman framework and how to inspire new algorithm classes with optimal approximations. In this paper, we present a novel topological viewpoint on temporal-difference (TD) based RL. We show that TD errors can be viewed as 1-cochain in the topological space of state transitions, while Markov dynamics are then interpreted as topological integrability. This novel view enables us to obtain a Hodge-type decomposition of TD errors into an integrable component and a topological residual, through a Bellman-de Rham projection. We further propose HodgeFlow Policy Search (HFPS) by fitting a potential network to minimize the non-integrable projection residual in RL, achieving stability/sensitivity guarantees. In numerical evaluations, HFPS is shown to significantly improve RL performance under non-Markovian.

The classification performance of deep neural networks relies strongly on access to large, accurately annotated datasets. In medical imaging, however, obtaining such datasets is particularly challenging since annotations must be provided by specialized physicians, which severely limits the pool of annotators. Furthermore, class boundaries can often be ambiguous or difficult to define which further complicates machine learning-based classification. In this paper, we want to address this problem and introduce a framework for mislabel detection in medical datasets. This is validated on the two largest, publicly available datasets for Video Capsule Endoscopy, an important imaging procedure for examining the gastrointestinal tract based on a video stream of lowresolution images. In addition, potentially mislabeled samples identified by our pipeline were reviewed and re-annotated by three experienced gastroenterologists. Our results show that the proposed framework successfully detects incorrectly labeled data and results in an improved anomaly detection performance after cleaning the datasets compared to current baselines.

Realistic sound propagation is essential for immersion in a virtual scene, yet physically accurate wave-based simulations remain computationally prohibitive for real-time applications. Wave coding methods address this limitation by precomputing and compressing impulse responses of a given scene into a set of scalar acoustic parameters, which can reach unmanageable sizes in large environments with many source-receiver pairs. We introduce Reciprocal Latent Fields (RLF), a memory-efficient framework for encoding and predicting these acoustic parameters. The RLF framework employs a volumetric grid of trainable latent embeddings decoded with a symmetric function, ensuring acoustic reciprocity. We study a variety of decoders and show that leveraging Riemannian metric learning leads to a better reproduction of acoustic phenomena in complex scenes. Experimental validation demonstrates that RLF maintains replication quality while reducing the memory footprint by several orders of magnitude. Furthermore, a MUSHRA-like subjective listening test indicates that sound rendered via RLF is perceptually indistinguishable from ground-truth simulations.

Sequential recommendation (SR) models predict a user's next interaction by modeling their historical behaviors. Transformer-based SR methods, notably BERT4Rec, effectively capture these patterns but incur significant computational overhead due to extensive intermediate computations associated with Softmax-based attention. We propose Cotten4Rec, a novel SR model utilizing linear-time cosine similarity attention, implemented through a single optimized compute unified device architecture (CUDA) kernel. By minimizing intermediate buffers and kernel-launch overhead, Cotten4Rec substantially reduces resource usage compared to BERT4Rec and the linear-attention baseline, LinRec, especially for datasets with moderate sequence lengths and vocabulary sizes. Evaluations across three benchmark datasets confirm that Cotten4Rec achieves considerable reductions in memory and runtime with minimal compromise in recommendation accuracy, demonstrating Cotten4Rec's viability as an efficient alternative for practical, large-scale sequential recommendation scenarios where computational resources are critical.

Grassroots Logic Programs (GLP) is a concurrent logic programming language with variables partitioned into paired \emph{readers} and \emph{writers}, conjuring both linear logic and futures/promises: an assignment is produced at most once via the sole occurrence of a writer (promise) and consumed at most once via the sole occurrence of its paired reader (future), and may contain additional readers and/or writers, enabling the concise expression of rich multidirectional communication modalities. GLP was designed as a language for grassroots platforms -- distributed systems with multiple instances that can operate independently of each other and of any global resource, and can coalesce into ever larger instances -- with its target architecture being smartphones communicating peer-to-peer. The operational semantics of Concurrent (single-agent) GLP and of multiagent GLP (maGLP) were defined via transition systems/multiagent transition systems, respectively. Here, we describe the mathematics developed to facilitate the workstation- and smartphone-based implementations of GLP by AI in Dart. We developed dGLP -- implementation-ready deterministic operational semantics for single-agent GLP -- and proved it correct with respect to the Concurrent GLP operational semantics; dGLP was used by AI as a formal spec, from which it developed a workstation-based implementation of GLP. We developed madGLP -- an implementation-ready multiagent operational semantics for maGLP -- and proved it correct with respect to the maGLP operational semantics; madGLP is deterministic at the agent level (not at the system level due to communication asynchrony), and is being used by AI as a formal spec from which it develops a smartphone-based implementation of maGLP.

Speculative decoding can significantly accelerate LLM serving, yet most deployments today disentangle speculator training from serving, treating speculator training as a standalone offline modeling problem. We show that this decoupled formulation introduces substantial deployment and adaptation lag: (1) high time-to-serve, since a speculator must be trained offline for a considerable period before deployment; (2) delayed utility feedback, since the true end-to-end decoding speedup is only known after training and cannot be inferred reliably from acceptance rate alone due to model-architecture and system-level overheads; and (3) domain-drift degradation, as the target model is repurposed to new domains and the speculator becomes stale and less effective. To address these issues, we present Aurora, a unified training-serving system that closes the loop by continuously learning a speculator directly from live inference traces. Aurora reframes online speculator learning as an asynchronous reinforcement-learning problem: accepted tokens provide positive feedback, while rejected speculator proposals provide implicit negative feedback that we exploit to improve sample efficiency. Our design integrates an SGLang-based inference server with an asynchronous training server, enabling hot-swapped speculator updates without service interruption. Crucially, Aurora supports day-0 deployment: a speculator can be served immediately and rapidly adapted to live traffic, improving system performance while providing immediate utility feedback. Across experiments, Aurora achieves a 1.5x day-0 speedup on recently released frontier models (e.g., MiniMax M2.1 229B and Qwen3-Coder-Next 80B). Aurora also adapts effectively to distribution shifts in user traffic, delivering an additional 1.25x speedup over a well-trained but static speculator on widely used models (e.g., Qwen3 and Llama3).

We study off-policy reinforcement learning for controlling continuous-time Markov diffusion processes with discrete-time observations and actions. We consider model-free algorithms with function approximation that learn value and advantage functions directly from data, without unrealistic structural assumptions on the dynamics. Leveraging the ellipticity of the diffusions, we establish a new class of Hilbert-space positive definiteness and boundedness properties for the Bellman operators. Based on these properties, we propose the Sobolev-prox fitted $q$-learning algorithm, which learns value and advantage functions by iteratively solving least-squares regression problems. We derive oracle inequalities for the estimation error, governed by (i) the best approximation error of the function classes, (ii) their localized complexity, (iii) exponentially decaying optimization error, and (iv) numerical discretization error. These results identify ellipticity as a key structural property that renders reinforcement learning with function approximation for Markov diffusions no harder than supervised learning.

Lewis' account of common knowledge in Convention describes the generation of higher-order expectations between agents as hinging upon agents' inductive standards and a shared witness. This paper attempts to draw from insights in learning theory to provide a formal account of common inductive knowledge and how it can be generated by a witness. Our language has a rather rich syntax in order to capture equally rich notions central to Lewis' account of common knowledge; for instance, we speak of an agent 'having some reason to believe' a proposition and one proposition 'indicating' to an agent that another proposition holds. A similar line of work was pursued by Cubitt & Sugden 2003; however, their account was left wanting for a corresponding semantics. Our syntax affords a novel topological semantics which, following Kelly 1996's approach in The Logic of Reliable Inquiry, takes as primitives agents' information bases. In particular, we endow each agent with a 'switching tolerance' meant to represent their personal inductive standards for learning. Curiously, when all agents are truly inductive learners (not choosing to believe only those propositions which are deductively verified), we show that the set of worlds where a proposition $P$ is common inductive knowledge is invariant of agents' switching tolerances. Contrarily, the question of whether a specific witness $W$ generates common inductive knowledge of $P$ is sensitive to changing agents' switching tolerances. After establishing soundness of our proof system with respect to this semantics, we conclude by applying our logic to solve an 'inductive' variant of the coordinated attack problem.

Autonomous drone racing pushes the boundaries of high-speed motion planning and multi-agent strategic decision-making. Success in this domain requires drones not only to navigate at their limits but also to anticipate and counteract competitors' actions. In this paper, we study a fundamental question that arises in this domain: how deeply should an agent strategize before taking an action? To this end, we compare two planning paradigms: the Model Predictive Game (MPG), which finds interaction-aware strategies at the expense of longer computation times, and contouring Model Predictive Control (MPC), which computes strategies rapidly but does not reason about interactions. We perform extensive experiments to study this trade-off, revealing that MPG outperforms MPC at moderate velocities but loses its advantage at higher speeds due to latency. To address this shortcoming, we propose a Learned Model Predictive Game (LMPG) approach that amortizes model predictive gameplay to reduce latency. In both simulation and hardware experiments, we benchmark our approach against MPG and MPC in head-to-head races, finding that LMPG outperforms both baselines.

Deep learning models trained to optimize average accuracy often exhibit systematic failures on particular subpopulations. In real world settings, the subpopulations most affected by such disparities are frequently unlabeled or unknown, thereby motivating the development of methods that are performant on sensitive subgroups without being pre-specified. However, existing group-robust methods typically assume prior knowledge of relevant subgroups, using group annotations for training or model selection. We propose Low-rank Error Informed Adaptation (LEIA), a simple two-stage method that improves group robustness by identifying a low-dimensional subspace in the representation space where model errors concentrate. LEIA restricts adaptation to this error-informed subspace via a low-rank adjustment to the classifier logits, directly targeting latent failure modes without modifying the backbone or requiring group labels. Using five real-world datasets, we analyze group robustness under three settings: (1) truly no knowledge of subgroup relevance, (2) partial knowledge of subgroup relevance, and (3) full knowledge of subgroup relevance. Across all settings, LEIA consistently improves worst-group performance while remaining fast, parameter-efficient, and robust to hyperparameter choice.

Can general-purpose AI architectures go beyond prediction to discover the physical laws governing the universe? True intelligence relies on "world models" -- causal abstractions that allow an agent to not only predict future states but understand the underlying governing dynamics. While previous "AI Physicist" approaches have successfully recovered such laws, they typically rely on strong, domain-specific priors that effectively "bake in" the physics. Conversely, Vafa et al. recently showed that generic Transformers fail to acquire these world models, achieving high predictive accuracy without capturing the underlying physical laws. We bridge this gap by systematically introducing three minimal inductive biases. We show that ensuring spatial smoothness (by formulating prediction as continuous regression) and stability (by training with noisy contexts to mitigate error accumulation) enables generic Transformers to surpass prior failures and learn a coherent Keplerian world model, successfully fitting ellipses to planetary trajectories. However, true physical insight requires a third bias: temporal locality. By restricting the attention window to the immediate past -- imposing the simple assumption that future states depend only on the local state rather than a complex history -- we force the model to abandon curve-fitting and discover Newtonian force representations. Our results demonstrate that simple architectural choices determine whether an AI becomes a curve-fitter or a physicist, marking a critical step toward automated scientific discovery.

Hallucinations in large language models remain a persistent challenge, particularly in multilingual and generative settings where factual consistency is difficult to maintain. While recent models show strong performance on English-centric benchmarks, their behavior across languages, tasks, and hallucination types is not yet well understood. In this work, we introduce Halluverse-M^3, a dataset designed to enable systematic analysis of hallucinations across multiple languages, multiple generation tasks, and multiple hallucination categories. Halluverse-M^3 covers four languages, English, Arabic, Hindi, and Turkish, and supports two generation tasks: question answering and dialogue summarization. The dataset explicitly distinguishes between entity-level, relation-level, and sentence-level hallucinations. Hallucinated outputs are constructed through a controlled editing process and validated by human annotators, ensuring clear alignment between original content and hallucinated generations. Using this dataset, we evaluate a diverse set of contemporary open-source and proprietary language models on fine-grained hallucination detection. Our results show that question answering is consistently easier than dialogue summarization, while sentence-level hallucinations remain challenging even for the strongest models. Performance is highest in English and degrades in lower-resource languages, with Hindi exhibiting the lowest detection accuracy. Overall, Halluverse-M^3 provides a realistic and challenging benchmark for studying hallucinations in multilingual, multi-task settings. We release the dataset to support future research on hallucination detection and mitigation\footnote{https://huggingface.co/datasets/sabdalja/HalluVerse-M3}.

The advancement of machine learning in audio analysis has opened new possibilities for technology-enhanced music education. This paper introduces a framework for automatic singing mistake detection in the context of music pedagogy, supported by a newly curated dataset. The dataset comprises synchronized teacher learner vocal recordings, with annotations marking different types of mistakes made by learners. Using this dataset, we develop different deep learning models for mistake detection and benchmark them. To compare the efficacy of mistake detection systems, a new evaluation methodology is proposed. Experiments indicate that the proposed learning-based methods are superior to rule-based methods. A systematic study of errors and a cross-teacher study reveal insights into music pedagogy that can be utilised for various music applications. This work sets out new directions of research in music pedagogy. The codes and dataset are publicly available.

Healthcare professionals (HCPs) face increasing occupational stress and burnout. Supporting HCPs need for relatedness is fundamental to their psychological wellbeing and resilience. However, how technologies could support HCPs relatedness in the workplace remains less explored. This study incorporated semi-structured interviews (n = 15) and co-design workshops (n = 21) with HCPs working in the UK National Health Service (NHS), to explore their current practices and preferences for workplace relatedness support, and how technology could be utilized to benefit relatedness. Qualitative analysis yielded a four-layer model of HCPs relatedness need, which includes Informal Interactions, Camaraderie and Bond, Community and Organizational Care, and Shared Identity. Workshops generated eight design concepts (e.g., Playful Encounter, Collocated Action, and Memories and Stories) that operationalize the four relatedness need layers. We conclude by highlighting the theoretical relevance, practical design implications, and the necessity to strengthen relatedness support for HCPs in the era of digitalization and artificial intelligence.

As AI systems increasingly become embedded in interactive and im-mersive artistic environments, artists and technologists are discovering new opportunities to engage with their interpretive and autonomous capacities as creative collaborators in live performance. The focus of this work-in-progress is on outlining conceptual and technical foundations under which performance-makers and interactive architecture can collaborate within rehearsal settings. It introduces a rehearsal-oriented prototype system for shaping and testing AI-mediated environments within creative practice. This approach treats interactive architecture as a performative agent that senses spatial behaviour and speech, interprets these signals through a large language model, and generates real-time environmental adaptations. Designed for deployment in physical performance spaces, the system employs virtual blueprints to support iterative experimentation and creative dialogue between artists and AI agents, using reasoning traces to inform architectural interaction design grounded in dramaturgical principles.

Vision capabilities in vision large language models (VLLMs) have consistently lagged behind their linguistic capabilities. In particular, numerous benchmark studies have demonstrated that VLLMs struggle when fine-grained visual information or spatial reasoning is required. However, we do not yet understand exactly why VLLMs struggle so much with these tasks relative to others. Some works have focused on visual redundancy as an explanation, where high-level visual information is uniformly spread across numerous tokens and specific, fine-grained visual information is discarded. In this work, we investigate this premise in greater detail, seeking to better understand exactly how various types of visual information are processed by the model and what types of visual information are discarded. To do so, we introduce a simple synthetic benchmark dataset that is specifically constructed to probe various visual features, along with a set of metrics for measuring visual redundancy, allowing us to better understand the nuances of their relationship. Then, we explore fine-tuning VLLMs on a number of complex visual tasks to better understand how redundancy and compression change based upon the complexity of the data that a model is trained on. We find that there is a connection between task complexity and visual compression, implying that having a sufficient ratio of high complexity visual data is crucial for altering the way that VLLMs distribute their visual representation and consequently improving their performance on complex visual tasks. We hope that this work will provide valuable insights for training the next generation of VLLMs.

Fully unsupervised segmentation pipelines naively seek the most salient object, should this be present. As a result, most of the methods reported in the literature deliver non-deterministic partitions that are sensitive to initialization, seed order, and threshold heuristics. We propose PANC, a weakly supervised spectral segmentation framework that uses a minimal set of annotated visual tokens to produce stable, controllable, and reproducible object masks. From the TokenCut approach, we augment the token-token affinity graph with a handful of priors coupled to anchor nodes. By manipulating the graph topology, we bias the spectral eigenspace toward partitions that are consistent with the annotations. Our approach preserves the global grouping enforced by dense self-supervised visual features, trading annotated tokens for significant gains in reproducibility, user control, and segmentation quality. Using 5 to 30 annotations per dataset, our training-free method achieves state-of-the-art performance among weakly and unsupervised approaches on standard benchmarks (e.g., DUTS-TE, ECSSD, MS COCO). Contrarily, it excels in domains where dense labels are costly or intra-class differences are subtle. We report strong and reliable results on homogeneous, fine-grained, and texture-limited domains, achieving 96.8% (+14.43% over SotA), 78.0% (+0.2%), and 78.8% (+0.37%) average mean intersection-over-union (mIoU) on CrackForest (CFD), CUB-200-2011, and HAM10000 datasets, respectively. For multi-object benchmarks, the framework showcases explicit, user-controllable semantic segmentation.

As increasingly capable open-weight large language models (LLMs) are deployed, improving their tamper resistance against unsafe modifications, whether accidental or intentional, becomes critical to minimize risks. However, there is no standard approach to evaluate tamper resistance. Varied data sets, metrics, and tampering configurations make it difficult to compare safety, utility, and robustness across different models and defenses. To this end, we introduce TamperBench, the first unified framework to systematically evaluate the tamper resistance of LLMs. TamperBench (i) curates a repository of state-of-the-art weight-space fine-tuning attacks and latent-space representation attacks; (ii) enables realistic adversarial evaluation through systematic hyperparameter sweeps per attack-model pair; and (iii) provides both safety and utility evaluations. TamperBench requires minimal additional code to specify any fine-tuning configuration, alignment-stage defense method, and metric suite while ensuring end-to-end reproducibility. We use TamperBench to evaluate 21 open-weight LLMs, including defense-augmented variants, across nine tampering threats using standardized safety and capability metrics with hyperparameter sweeps per model-attack pair. This yields novel insights, including effects of post-training on tamper resistance, that jailbreak-tuning is typically the most severe attack, and that Triplet emerges as a leading alignment-stage defense. Code is available at: https://github.com/criticalml-uw/TamperBench

The recent surge in Time Series Foundation Models has rapidly advanced the field, yet the heterogeneous training setups across studies make it difficult to attribute improvements to architectural innovations versus data engineering. In this work, we investigate the potential of a standard patch Transformer, demonstrating that this generic architecture achieves state-of-the-art zero-shot forecasting performance using a straightforward training protocol. We conduct a comprehensive ablation study that covers model scaling, data composition, and training techniques to isolate the essential ingredients for high performance. Our findings identify the key drivers of performance, while confirming that the generic architecture itself demonstrates excellent scalability. By strictly controlling these variables, we provide comprehensive empirical results on model scaling across multiple dimensions. We release our open-source model and detailed findings to establish a transparent, reproducible baseline for future research.

Background: Transcranial magnetic stimulation (TMS) is a powerful tool to investigate neurophysiology of the human brain and treat brain disorders. Traditionally, therapeutic TMS has been applied in a one-size-fits-all approach, disregarding inter- and intra-individual differences. Brain state-dependent EEG-TMS, such as coupling TMS with a pre-specified phase of the sensorimotor mu-rhythm, enables the induction of differential neuroplastic effects depending on the targeted phase. But this approach is still user-dependent as it requires defining an a-priori target phase. Objectives: To present a first realization of a machine-learning-based, closed-loop real-time EEG-TMS setup to identify user-independently the individual mu-rhythm phase associated with high- vs. low-corticospinal excitability states. Methods: We applied EEG-TMS to 25 participants targeting the supplementary motor area-primary motor cortex network and used a reinforcement learning algorithm to identify the mu-rhythm phase associated with high- vs. low corticospinal excitability. We employed linear mixed effects models and Bayesian analysis to determine effects of reinforced learning on corticospinal excitability indexed by motor evoked potential amplitude, and functional connectivity indexed by the imaginary part of resting-state EEG coherence. Results: Reinforcement learning effectively identified the mu-rhythm phase associated with high- vs. low-excitability states, and their repetitive stimulation resulted in long-term increases vs. decreases in functional connectivity in the stimulated sensorimotor network. Conclusions: We demonstrated for the first time the feasibility of closed-loop EEG-TMS in humans, a critical step towards individualized treatment of brain disorders.

Structural parameters of graphs, such as treewidth, play a central role in the study of the parameterized complexity of graph problems. Motivated by the study of parametrized algorithms on phylogenetic networks, scanwidth was introduced recently as a new treewidth-like structural parameter for directed acyclic graphs (DAGs) that respects the edge directions in the DAG. The utility of this width measure has been demonstrated by results that show that a number of problems that are fixed-parameter tractable (FPT) with respect to both treewidth and scanwidth allow algorithms with a better dependence on scanwidth than on treewidth. More importantly, these scanwidth-based algorithms are often much simpler than their treewidth-based counterparts: the name ``scanwidth'' reflects that traversing a tree extension (the scanwidth-equivalent of a tree decomposition) of a DAG amounts to ``scanning'' the DAG according to a well-chosen topological ordering. While these results show that scanwidth is useful especially for solving problems on phylogenetic networks, all problems studied through the lens of scanwidth so far are either FPT with respect to both scanwidth and treewidth, or W[$\ell$]-hard, for some $\ell \ge 1$, with respect to both. In this paper, we show that scanwidth is not just a proxy for treewidth and provides information about the structure of the input graph not provided by treewidth, by proving a fairly stark complexity-theoretic separation between these two width measures. Specifically, we prove that Weighted Phylogenetic Diversity with Dependencies is FPT with respect to the scanwidth of the food web but W[$\ell$]-hard with respect to its treewidth, for all $\ell \ge 1$. To the best of our knowledge, no such separation between these two width measures has been shown for any problem before.