Papers

As large language model (LLM) agents become widely adopted, they are increasingly deployed for tasks that require persistent monitoring or recurring actions (e.g., market analysis). These agents are expected to operate unattended for days or weeks, act at the right timing, and adapt to the dynamic environment over time. These challenges are not fully captured in the existing long-horizon agent work, as they often consider a static environment that is not temporally changing. We introduce ReLiveGym, a diagnostic evaluation environment of long-lived tasks in which agents act sparsely over simulated weeks of chronologically replayed real-world news, market, and social-media streams. The tasks span diverse levels of time sensitivity, reasoning intensity, and recurrence. Across eight base language models, we investigate how model choice and harness design affect agent performance on such long-lived tasks. Our results show that how agents determine when to act arises as an important harness-design axis for long-lived tasks; and that the optimal design varies across tasks and sometimes model choices as well. We also evaluate how continuous learning from hindsight feedback affects performance and addresses failure modes observed in these long-lived tasks. These findings indicate model choice, action timing mechanism, and use of feedback as important considerations in the design of long-lived agents. Code: https://github.com/SaharaLabsAI/ReLiveGym

Data-driven Riemannian geometry provides nonlinear interpolation and geometric representations of high-dimensional data. For these operations to be statistically meaningful, paths between observations should preferentially traverse high-likelihood regions. Existing scalable pullback constructions typically use a unimodal Gaussian latent distribution, assuming that the data reside close to a single manifold. For multimodal data, mapping separated modes or local structures into one Gaussian region can require substantial transport deformation and compromise the resulting geometry. We introduce a pullback geometry for data supported on mixtures of manifolds. Using a latent Gaussian mixture, we define its Riemannian metric as the matrix square of the responsibility-weighted expected component precision. The metric is smooth and positive definite and recovers the existing Gaussian construction in the single-component limit. For structured overlapping mixtures, we establish conditions under which the log-density is concave along geodesics, providing a formal connection between the proposed geometry and paths through high-likelihood regions, and derive the corresponding local curvature relations. We instantiate this geometry in a normalizing flow with adaptive mixture learning, allowing the number of active components to emerge from the data and supporting component-wise reconstruction and local effective-dimension estimation. Experiments on synthetic geometric data, a controlled multi-view image setting with a known reference trajectory, and MNIST show reduced transport distortion, competitive path support, close reference-trajectory recovery, and improved interpolation realism. These results extend scalable pullback geometry beyond datasets that reside close to a single manifold while retaining tractable and interpretable local structure.

Large Language Models (LLMs) have been used for novel discovery of algorithms, theorems, drugs, and other tasks through the use of harnesses that prompt an LLM to iteratively optimize an objective. In this work, we study the relationship between the population of previous iterates and eventual discovery success. We generalize past work on harness design to develop a suite of 12 harnesses called 'Modular' and characterize their performance across 5 diverse discovery tasks, finding that discovery success is brittle and sensitive to harness design. We uncover mode collapse, characterized by a dramatic drop in the diversity of iterates, as a common failure mode. We find that popular state-of-the-art harnesses and diversity-inducing harness interventions, which aim to prolong this collapse, yield inconsistent gains. Our results instead uncover that the performance of early discoveries is predictive of eventual success. We therefore propose a universally applicable intervention that performs an initial stage of parallel exploration in order to initialize subsequent iterative optimization. Our method provides consistent gains across many harnesses and target applications, confirming the importance of initialization in LLM-driven discovery.

Large audio-language models (LALMs) are sensitive to input perturbations, such as noise, waveform corruption, and adversarial injections. We propose AnchorPrompt, an efficient adaptation method that keeps the model frozen and learns a single block of prompt vectors inserted at the decoder input, between the audio and question embeddings. We train these vectors through self-distillation over diverse audio and text perturbations. To improve answer consistency and mitigate hallucination, we use the model's prediction on the clean recording as the target for answerable inputs, and assign a refusal target when the audio lacks sufficient evidence to answer. Furthermore, AnchorPrompt is perturbation-agnostic at inference, requiring no prior detection of perturbations and enabling zero-shot transfer to unseen distortions. We evaluate three LALMs across three benchmarks and show that AnchorPrompt improves answer consistency in most tested conditions. Clean accuracy improves in six of nine model-benchmark pairs, with minimal impact on the remainder of 1.2% at most. Crucially, AnchorPrompt reduces hallucinations under severe audio corruption while keeping false refusals on clean audio rare. Finally, these consistency gains transfer to unseen perturbations, such as choice permutations and reverberation.

This paper develops a meta-multi-agent reinforcement learning (meta-MARL) framework to enable fast adaptation of interactive policies in a multi-agent system (MAS). Meta-reinforcement learning (meta-RL) enables agents to rapidly adapt to new tasks/environments using a bi-level optimization mechanism. However, existing meta-RL generally focuses on single-agent systems. Extending these frameworks and algorithms to multi-agent systems poses additional challenges, as tasks are characterized by not only the environment but also agents' strategic interactions. To address these challenges, we model multi-agent reinforcement learning (MARL) problems as Markov games (MGs) and develop a meta-MARL framework for rapid interactive policy adaptation across a distribution of MGs. A new concept, called meta-NE, is defined to describe the desired solution concept in a meta-MARL problem. Sufficient conditions for the equivalence between a meta-NE and a stationary point of the gradient-play-based meta-MARL algorithm are established. Our evaluation on autonomous-driving tasks demonstrates that the proposed meta-MARL method achieves faster adaptation than pretrained MARL baselines, validating the effectiveness of our framework.

Natural-language skills are textual procedural memories through which large language model (LLM) agents retain reusable task knowledge without updating model weights. Existing methods typically treat skills as either static artifacts or monolithic documents optimized using aggregate validation scores as feedback. However, representing a skill as a monolithic document restricts optimization to its textual content, without explicitly modeling the structure through which procedural knowledge is retrieved and executed. We identify a key distinction between learning what knowledge to retain and determining how to organize it: textual updates should first be validated through execution evidence, after which the retained knowledge should be structured according to its procedural dependencies and retrieval requirements. To this end, we introduce SkillSpec, a two-phase framework comprising consensus-gated evolution and representation specialization. In the consensus-gated phase, complementary editing intents generate complete candidate skills. An update is committed only when paired evaluations reach consensus, requiring sufficient overall improvement and non-negative aggregate paired gain in every repeated evaluation. In the specialization phase, signals of process and redundancy sensitivity derived from the full optimization trajectory, including accepted and rejected candidates, guide the selection of a flat, graph, or hybrid representation.Across six benchmarks and three target language models, SkillSpec improves average success rate over SkillOpt by 6.89%, averaged across the three models. These results demonstrate that reliable skill evolution and representation specialization address complementary objectives: deciding what knowledge to retain and how to structure it for inference.

Linearizability gives developers the illusion that operations on a distributed datastore execute sequentially on a single machine. Providing it is expensive, so numerous specialized reads algo- rithms have been proposed to make reads significantly faster without sacrificing linearizability, which matters because most workloads are read-dominant by orders of magnitude. Yet, none of these algorithms is the best choice for every deployment. We show why with a framework that classifies these algorithms into a generalized space together with a mathematical analysis of their latencies as a function of this generalization. The analysis exposes two fundamental tradeoffs, one between read and write latency and one between read and write tolerance to network variance, which together leave the best algorithm dependent on network and workload conditions that vary over time. It also exposes a gap: there is a way to minimize the delay a read incurs when it overlaps a concurrent write, at no cost to writes, yet exist- ing algorithms can apply it only for a subset of the tradeoff space, forcing any deployment that needs the full space to settle for a strictly worse solution. We close this gap with a new algorithm, Pairwise Quorums, and build a second, Ditto, that reconfigures across the space at runtime as conditions change. Together they subsume every existing algorithm under our model: against each, the combination is either strictly better, or reconfigures to match it where a genuine tradeoff applies.

Multimodal large language models (LLMs) increasingly integrate vision and text, yet how people use them in natural settings remains underexplored. We seek to answer the question: when users upload images, what tasks are they trying to accomplish? Analyzing over 40,000 de-identified image-upload conversations from Microsoft Copilot, we characterize real-world multimodal use through a hierarchical framework of ten capabilities, spanning perception, cognition, and generation. First, we characterize the distribution and composition of these capabilities, finding that the majority of image-upload tasks involve multiple capabilities. Second, we find that multimodal use spans a broader and more diverse task space than text-only interactions, with asymmetric coverage and task classes that rely on cross-modal grounding. Third, mapping observed capability demand onto 253 existing benchmarks reveals uneven alignment between benchmark coverage and real-world use: benchmarks concentrate on perception and reasoning toward fixed answers, while common workflows involving text, code, and data generation remain comparatively undertested. We validate our taxonomy and findings on an independent ChatGPT dataset. Our results provide a large-scale empirical characterization of what users seek to accomplish with multimodal LLMs and highlight opportunities for benchmark design grounded in observed user demand.

Recent advances in AI for scientific discovery enable molecular understandingand design, yet reasoning over incomplete chemical representations remainsunclear.Markush structures, which encode molecular families through variable R-groupplaceholders (\textit{R\textsubscript{1}}, \textit{R\textsubscript{2}}, X, etc.), are ubiquitous in pharmaceutical patents and requiregrounding across molecular, textual, and chemical information.However, existing molecule-language benchmarks focus on fully specifiedmolecules, leaving R-group grounding largely unevaluated.We introduce R-GroundBench:, a diagnostic benchmark built from real patent Markushstructures, featuring a Multiple-Choice (VQA) track with controlled difficultyand modality splits, and an open-ended Generation track.Our results reveal a substantial gap between recognition andmolecular grounding.While models achieve over 90% accuracy on Easy VQA, performance drops to56–66% on Hard VQA when shortcuts are controlled.Chemical-domain VLMs also remain unreliable, achieving only 25.7–46.2% on HardVQA despite domain-specific pretraining.Moreover, Generation Exact Match remains below 20% for most models and below8% when visual input is required.These findings reveal that current AI systems lack reliable grounding andexecution for Markush editing, highlighting challenges for AI-drivenscientific discovery.

We develop multifidelity methods for constructing triangular transport maps from samples, when high-fidelity data are scarce but lower-fidelity data are more abundant. Using this set of multifidelity data, we approximate a triangular transport map that bijectively maps between a tractable reference density and the high-fidelity target distribution. We introduce two strategies to leverage low-fidelity data: a hierarchical approach that composes maps between adjacent fidelity levels, and a non-hierarchical method that incorporates low-fidelity information through monotonicity-preserving corrections to the map parameterization. Numerical experiments compare these strategies with single-fidelity transport and demonstrate how the proposed multifidelity approaches can improve map estimation from limited high-fidelity data. To illustrate the broader utility of the learned maps, we also deploy them in a downstream amortized simulation-based inference task. This example shows that multifidelity improvements in map estimation can translate to improved conditional sampling and uncertainty quantification when high-fidelity data are scarce.

We study bandwidth fee mechanisms (BFMs) for pricing threshold-certified transaction dissemination before consensus. Analogous to transaction fee mechanisms for computation, BFMs price blockchain communication, which is especially relevant when dissemination is separated from consensus and execution. We model BFMs as two-sided procurement auctions between users submitting transactions and validators receiving and forwarding them. Any validator may serve as a transaction's source and forward it to a threshold of attesters. The model accounts for receipt from the user and subsequent transfers between validators, while abstracting away unrestricted multi-hop gossip and network topology. We establish limits on myopic incentive compatibility: no nontrivial transcript-based BFM prevents profitable collusion between two validators, and no nontrivial BFM simultaneously satisfies incentive compatibility for individual users, individual validators, and coalitions of one validator and one user. On the feasibility side, we give a greedy-route posted-price mechanism that is incentive compatible for every individual user and validator. We also study bandwidth pricing across blocks. Assuming honest validator execution, we propose an EIP-1559-style price-update rule whose fluid approximation has a unique fixed point that is locally stable for sufficiently small step sizes. In large markets, prices starting near this point remain nearby with high probability over any fixed time horizon.

Secure aggregation lets a server recover an aggregate of client updates without observing any individual update, but conventional protocols fix the aggregate precision when clients upload. We introduce and formulate a new progressive-resolution secure-aggregation functionality in which clients upload once and successively finer resolutions of the same aggregate can later be authorized without renewed client participation. To realize this functionality, we propose progressive-resolution secure aggregation (PSA): each clipped, dithered update is represented by compatible nested-lattice digits; separately releasable layers are protected by secure aggregation and an additional aggregate pad that remains unavailable to the server until a non-colluding release controller authorizes that layer.

Compression reports summarize how far a compressed language model moved from the dense one, usually by a KL divergence; a deployment that relies on the dense model's outputs needs to know how many of its decisions changed. We show that total variation, not KL, answers this directly. Across 802 compressed and perturbed copies of 19 open models on five corpora and nine mechanically unrelated perturbation families, the rate at which the arg-max token changes (the flip rate) tracks total variation at a ratio with median $1.05$, with no fitted constant. KL converts into flips only through its square root and a factor that varies fourfold across models and corpora, because KL averages over tokens before the root is taken; first-order statistics averaged per token, such as Hellinger distance, avoid this, but reports rarely give them. As a result, of two compressors reported on different models and corpora whose flip rates differ by at least $10%$, KL assigns the smaller divergence to the one that changes more decisions in $11%$ of cases, total variation in $1%$. Two pre-registered tests mark the limits: on a held-out code corpus the ratio held for all eight models while three predictions about KL each failed for half of them or more, and on three new models with real kernels it stayed in its band for 37 of 38 checkpoints but fell below one on code for two models. In vLLM speculative decoding, total variation measured under teacher forcing predicts greedy draft acceptance with a mean relative error of $1.1$–$2.4%$, without the task-specific calibration that KL needs.

Federated learning (FL) has recently attracted increasing attention in remote sensing (RS) since it enables collaborative model training across decentralized RS image archives without requiring direct access to local data. However, FL performance significantly degrades when the data distributions between clients are heterogeneous, which often occurs due to geographical differences, seasonal changes, and varying image acquisition and atmospheric conditions. To address this challenge, in this letter, we propose a novel personalized FL framework (denoted as FedMAD) for RS image classification problems. The proposed framework separates globally shared representation parameters from client-specific adaptation parameters to preserve client-specific features while maintaining globally transferable representations. This is achieved by integrating lightweight modulation modules and local batch normalization layers into the backbone network. Although globally shared parameters are collaboratively optimized between clients, client-specific parameters remain local to preserve domain-specific feature characteristics. In addition, FedMAD introduces a modulation-aware directional aggregation strategy that dynamically adjusts the importance of aggregation for each client according to the alignment of local modulation updates. This allows the global optimization process to suppress conflicting client updates originating from heterogeneous data distributions while enhancing the contribution of clients with consistent adaptation behaviors. The experimental results obtained on the BigEarthNet-S2 and EuroSAT datasets demonstrate the effectiveness of FedMAD compared to state-of-the-art FL algorithms under heterogeneous RS data distributions. The code of the proposed framework will be publicly available at https://git.tu-berlin.de/rsim/fedmad.

Internationally mobile scientists are central to national research and innovation systems. We link faculty rosters from the Academic Analytics Research Center to OpenAlex publication records for 2011-2020, yielding more than 12 million faculty-publication observations for 236,394 tenure-system faculty at more than 300 major U.S. universities. Foreign-trained faculty, defined by a terminal degree awarded outside the United States, constitute about 11% of the observed faculty workforce but account for 13-14% of publications and 14-16% of top-1% cited elite output. We find that this elevated representation in elite output is closely related to their organizational embeddedness: when faculty are compared within the same institution, scientific domain, rank, and year, the difference in elite output narrow substantially while overall productivity and collaboration differences persist within those settings. In addition, foreign-trained faculty enter each publication year with larger and broader prior collaboration networks, and collaborator reach is more strongly associated with subsequent elite output. By contrast, although raw topic breadth is greater among foreign-trained faculty, after accounting for prior publication volume, we find that their topic breadth is slightly narrower and more cognitively concentrated. These findings recast international training in the lens of scientific capacity and organizational integration in that internationally accumulated scientific capabilities become embedded in institutions and relationships through which research is organized and produced.

Recent joint audio-video generative models can synthesize realistic videos with synchronized sound, but typically generate audio as a single mixed track. This limits source-level control and differs from practical audiovisual workflows, where speech, music, sound effects, and ambient sounds are represented as separate editable tracks. We introduce Soundwich, a training-free framework that transforms a frozen joint audio-video flow-matching model into a generator of multiple synchronized, independently editable audio stems coupled to a shared video. Soundwich generates separate audio stems with explicit control over their temporal activity. To keep separately generated sounds coherent, we introduce a shared scene representation that communicates global audiovisual context across stems while preserving their source-level separation. We further route cross-modal interactions between each audio stem and its corresponding visual source, improving audiovisual consistency. The resulting stems remain synchronized with the video and can be independently retimed, muted, replaced, or remixed. Experiments and human evaluations show improved temporal control, source separation, and naturalness, while enabling flexible source-level editing within coherent audiovisual generation. Code is available at https://github.com/CodyNing/Soundwich.

Large language models (LLMs) are widely used for claim verification, yet remain brittle for numerical reasoning: even small changes in value can sharply degrade accuracy. We show that this brittleness persists in frontier LLMs, but can be mitigated through adversarial fine-tuning on numerically perturbed examples. Using parameter-efficient fine-tuning, small Qwen3 models (0.6B$\unicode{x2013}$8B) reach 98.7% accuracy on label-flipping perturbations, outperforming larger zero-shot models and frontier systems (GPT-5.4 Pro (74.0%) and Gemini 2.5 Flash (73.9%)). The gains generalise to unseen perturbation types, indicating robust numerical decision boundaries rather than memorised edits. Robustness also transfers without target-domain data, significantly improving cross-lingual performance in Spanish. We further show that the same fine-tuning recipe confers robustness to evidence-side perturbations, using the VitaminC dataset.

This paper considers the following basic question: If an (ordered) linear-probing hash table is allowed two hash functions, instead of one, how does this change the expected insertion and query time, as a function of the load factor $1 - \epsilon$? We prove that the greedy two-choice insertion strategy achieves polynomially better bounds than the single choice algorithm, but that one can even do much better by using more sophisticated non-greedy strategies. Specifically, we show that there is an insertion strategy that does not evict elements (once an element is inserted, its hash choice is fixed) and that achieves expected query time $O(\log \epsilon^{-1})$ with expected insertion time $O(\epsilon^{-1})$. We then further show that, if one is allowed to evict elements (i.e., to change over time which hash function a given element uses), then it is possible to achieve expected query time $O(1)$ with expected insertion time $O(\epsilon^{-1/2})$. This final result achieves an expected query time of $O(1)$ even when the hash table is filled to $100\%$ full. Combined, the results reveal that there is a surprisingly strong “power of two choices” phenomenon for linear-probing hash tables, allowing for a two-choice hash table to achieve significantly better bounds than what might at first seem to be possible.

Hardware-operable failures (HOFs) interrupt large language model (LLM) training but permit recovery on the same hardware without reset, repair, or replacement. Existing recovery systems nevertheless reload checkpoints, recompute lost progress, and rebuild process state, idling GPUs that could otherwise continue training. We present Leto, a fault-tolerant training system that leverages surviving hardware to enable efficient in-place recovery. Our key insight is that the state needed to resume training can be retained or prepared outside the active training process while remaining on the same hardware. Leto retains the working model state and the reusable process state, and preinitializes the remaining state in a shadow trainer. We devise two-tier erasure protection and chunk-level transactional updates to keep the retained model state recoverable and consistent, and reclaim the shadow state when active training needs its GPU memory. Evaluation on 6- and 72-GPU NVIDIA A100 clusters shows that Leto recovers 3.6–6.5$\times$ faster than the best-performing checkpointing baselines and improves productive training time by up to 13.7 percentage points. Large-scale simulation shows over 95% productive training time on a 131,072-GPU cluster.

Recent video-based world models pair the scalability of autoregressive (AR) prediction with the visual quality of diffusion models. The choice of scene tokenizer is paramount for the optimal performance of each of these, both in terms of fidelity and semantics. Flexible-length, coarse-to-fine tokenizers yield exactly that: the first coarse tokens carry the clip's global semantics while later tokens further specify details. Existing flexible tokenizers only apply a representation-alignment (REPA) loss on early decoder hidden states, a target the decoder can partly meet from its noised input instead. We introduce SemanTok, a flexible video tokenizer that feeds frozen DINO features into its encoder and adds lightweight heads that reconstruct them from each retained token prefix alone. SemanTok achieves high semantic alignment and video fidelity at every AR model size: a 201M SemanTok AR model matches or beats a VideoFlexTok AR model $3.4\times$ its size, and larger SemanTok AR models further improve fidelity. It keeps semantic alignment on out-of-distribution classes and gives the decoder higher semantic alignment at every noise level, including pure noise. It performs well in both reconstruction and generation, and its short token prefixes are cheaper to predict and give better generation fidelity, with pixel detail deferred to later tokens.

With the rapid adoption of large language models (LLMs) and parameter-efficient fine-tuning (PEFT) methods, the risk of backdoor attacks has become more severe. Existing backdoor purification methods typically rely on at least one of the strong assumptions, such as prior knowledge of triggers, access to clean references, or aggressive retraining, and they often lack comprehensive evaluations. These constraints substantially limit their practical applicability. To overcome these challenges, our work proposes purifying LoRA-tuned LLMs without these assumptions and even without post-hoc retraining of the suspect parameters. Our objective is to significantly reduce the attack success rates (ASR) while preserving both (i) the base model's general capabilities and (ii) the new downstream skills learned through the adapter. Through a series of ablation studies, we progressively scale our approach from a single layer in a text classification setting to a full-parameter LLM in the generative task. Through careful data curation and feature approximation, we extract high-fidelity backdoor directions and, for each layer or head, construct orthogonal null spaces in both the input and output channels, onto which the LoRA updates are projected. Empirically, our null-space projection method reduces the ASR from nearly 100% to less than 10%, while preserving the base model's benign performance and the adapter's learned abilities during downstream task adaptation.

Companies and brands increasingly use dynamic pricing, including targeting social media ads to users based on their income. In this work we audit TikTok's feed using 56 automated accounts, which collect data on over 80,000 videos, across two studies. We test the impact of device price on ad load and ad types, using 12 phones of low ($0-$250), medium ($400-$650), and high ($750-$1,000+) price as a proxy for income. We also test the impact of interaction type (i.e., like, comment, share), age, and gender on the frequency and content of ads. Overall, the ad load was 29.4%, but liking and sharing content increased ad load significantly. We also found that as accounts spend more time on TikTok, the ad load steadily increases. We found some evidence that device price impacts both ad load and content – more expensive devices were targeted with fewer ads, while the least expensive devices received more discounts.

We introduce Grand Canonical Generators (GCG), a generative framework that extends Boltzmann generators to the grand canonical ensemble. We present two designs. The first conditions a variable-size generative model on the chemical potential, sampling particle number and configuration jointly. The second factorizes the grand canonical distribution into a particle-number distribution and the corresponding canonical Boltzmann density. This factorized formulation can use any existing Boltzmann generator for the canonical component, encodes the known linear chemical-potential dependence analytically, and yields a tractable likelihood that supports self-normalized importance sampling (SNIS). Empirically, GCG accurately reproduces grand canonical observables on a Lennard–Jones fluid and methane adsorption in a zeolite, demonstrating generalization across chemical potentials and correction via SNIS and grand canonical Monte Carlo.

Large language models (LLMs) increasingly underpin scientific AI applications that reason over structured knowledge, from biomedical question answering to materials informatics. However, their logical reasoning often falls short, producing factual inaccuracies unacceptable in these settings. Reliable evaluation remains challenging: manual dataset construction scales poorly, and LLM-based generation risks embedding the very flaws it aims to measure. High-quality benchmarks must ground both correct and incorrect labelled examples in explicit background knowledge, formally verifiable by a standard reasoner. We propose a pipeline that automatically generates ontology-grounded multiple-choice question (MCQ) benchmarks from any sufficiently axiomatised OWL 2 ontology, with correct answers grounded in the ontology by design. Distractors are generated by perturbing the right-hand-side class expressions of class definition axioms, and their incorrectness is formally verified by an OWL reasoner via entailment checks. We evaluate the pipeline on three ontologies: Pizza (small, academic), PMDco (complex, materials science), and DOID (large, biomedical), generating 112, 2,491, and 15,216 MCQs respectively. Distractors span four semantic categories from class unsatisfiability to weakened subsumptions, enabling diagnostic evaluation of specific reasoning failures. Items meet natural language quality standards: mean LLM judge scores of 4.02, 4.36, and 3.36 out of 5 confirm fluency, and correct-answer-to-distractor similarity above 0.8 shows that wrong options cannot be dismissed on surface form alone. Six LLMs evaluated zero-shot achieve 41.1-76.8% accuracy, well above the 25% random-guessing baseline, indicating the benchmarks are challenging and discriminative. This work is a step towards more reliable benchmarks for assessing logical reasoning in scientific AI.

Curvature is often treated as an intrinsic property of a representation, although its empirical effect also depends on coordinate scale, learned logit temperature, and numerical safeguards. We study this interaction in hZACH-ViT, a compact Vision Transformer with Euclidean, Poincare, and spherical prototype heads. The backbone architecture, seed-specific initialization, 50-per-class training subset, and optimization protocol are matched across three MedMNIST datasets and five seeds. At the fixed comparison curvature $c=1$, Poincare has the lowest class-macro PGD attack-success rate in all 12 dataset-budget cells and under a stronger CE+DLR multi-restart attack on all three datasets, but it also has the lowest clean MacroF1. An end-to-end curvature intervention changes the interpretation. Reducing Poincare curvature to $c=0.1$ improves clean MacroF1 in every one of the 15 paired seed-dataset comparisons and removes hard boundary clipping, yet on OrganAMNIST it increases strong attack success from $89.7\%$ to $99.3\%$ (paired difference $+9.57$ points; 95% hierarchical bootstrap CI $[+5.52,+14.03]$). At $c=1$, $40$-$47\%$ of clean Poincare features are hard-clipped, the radial Jacobian of the inherited map is nearly zero, and dimensionless attack trajectories are unusually long and inefficient. The spherical head provides a control: its curvature change is an exact scale gauge to floating-point precision and produces much smaller attack differences. These results do not establish intrinsic hyperbolic robustness. They identify an implementation-sensitive regime in which curvature, scale, and proximity to the Poincare boundary jointly organize clean recognition and adversarial representation motion.

Language models (LMs) are increasingly used for tasks that require reasoning about hidden variables from a few observations, for which Bayesian inference is the normatively correct solution. While supervised fine-tuning of an LM on the outputs of an optimal $\textit{Bayesian}$ model leads to near-Bayesian behavior, standard supervised fine-tuning (SFT) on the true answers for the task falls short of it. But behavior alone does not tell us $\textit{why}$ tuning on a $\textit{Bayesian}$ or an $\textit{oracle}$ (true answers) signal differs: whether the resulting LM represents Bayesian beliefs, acts on them, or turns them into a choice the way Bayes' rule does. To compare them, we formulate increasingly demanding requirements for an LM to count as a Bayesian decision maker, spanning its behavior, representations, and computations, and test them on a flight recommendation task. The Bayes-trained LM acts Bayesian, encodes quantities of Bayes' rule in its middle layers, and uses the encoded belief for the recommendation to a certain extent. The oracle-trained LM differs from it both in the beliefs it holds and whether it reads beliefs out into recommendations. Exchanging beliefs between the LMs transfers a part of the Bayesian advantage. Bayes fine-tuning thus installs usable Bayesian beliefs in an LM for reasoning under uncertainty in a way standard SFT on oracle answers cannot, highlighting the advantage of nuanced supervision.

Histological whole-slide images (WSIs) are central to computational pathology but pose severe computational challenges due to their extremely high resolution, often spanning several gigabytes per slide. To enable scalable learning, existing methods apply self-supervised data condensation to reduce computational cost, but typically rely on heuristic prototype learning and do not explicitly preserve learning-relevant feature distributions for downstream tasks. In response, we introduce a principled reformulation of WSI condensation as a distribution-matching problem under a fixed representational lens, and develop NICER, a tractable approximation framework based on a nonparametric prior with slide-adaptive capacity. Experiments on five histopathology datasets, together with clinical evaluation from a board-certified pathologist, show that NICER consistently outperforms prior methods, achieving an average accuracy improvement of 7.44% while offering improved efficiency-accuracy trade-offs, highlighting the benefits of principled, distribution-aware condensation for scalable histological representation learning. Source codes are available in https://github.com/nmduonggg/NICER.

Advancements in augmented reality (AR) continue to foster innovative solutions, facilitating novel methodologies within educational systems, healthcare delivery, and risk-mitigation protocols. However, optimizing for end-user immersion and comfort remains challenging, as AR head-mounted displays contend with constrained scene geometry, spatial jitter, and temporal instability. User studies are the standard AR evaluation method for visual quality, but their cost, diminishing scalability, and inflexibility pose bottlenecks during iterative application design. To address this problem, we present an automated framework for AR content evaluation and refinement, built on vision-language models (VLMs), to evaluate and predict the visual fidelity of AR scenes as perceived by users. First, we introduce RateAR, a benchmark of AR images and videos collected across diverse scenes and environmental conditions, with good-to-excellent reliability (ICC(2,5) >= .90) across perceptual factors, including object placement, scale, and shadow consistency. Subsequently, we evaluate eleven commercial VLMs on the crafted benchmark. Results support that VLM-based quality predictions strongly correlate with human subjective judgments, achieving Spearman's rank-order correlations of up to 0.8695. An ablation study further suggests that, compared to other prompting strategies, our contextual prompting yields better alignment with human ratings while balancing introduced complexity cues. Building on these findings, we construct an automated AR content adjustment system and conduct a 21-participant user study. More than 90% of participants found that the system improved placement and size coherence of virtual content.

Unsupervised skill discovery has emerged as a promising approach for leveraging reward-free datasets to pretrain general-purpose policies. However, current skill discovery methods either require access to expert data or exhibit limited generalization, failing to transfer effectively to previously unseen layouts. A key challenge is to learn representations that capture the temporal structure of the environment while remaining robust to variations across layouts. To address this issue, we present an objective for learning action-aware temporal representations that satisfy the functional equivariance property while preserving the local temporal structure of the environment. Building upon this embedding, we further propose unsupervised skill discovery using bisimulation, which learns transferable skills by conditioning the behavior of skills exclusively on the subset of state features that directly affect their execution. This enforces invariant behavior across different layouts, enabling skills to transfer effectively to other configurations. Finally, through comprehensive empirical evaluations, we show that skills learned in a given environment can be effectively applied to solve downstream tasks in various environment layouts, demonstrating strong out-of-distribution generalization.

Evolutionary approaches to LLM-driven discovery often generate new programs from a small set of selected ancestors. This keeps contexts manageable but can omit useful evidence from other experiments, whereas including the full experimental history produces long, redundant contexts. We introduce a simple, single-agent discovery harness built around LabBook, an agent-maintained memory that serves two complementary roles: guiding retrieval of relevant evidence from a complete experimental log and informing the generation of new solutions. At each iteration, the same agent combines its memory with retrieved evidence and jointly produces the next program and an updated LabBook. This separates complete history retention from selective context construction, without requiring an explicit population or branching search structure. On 49 Frontier-CS problems, LabBook improves the observed quality-cost trade-off over the evaluated evolutionary baselines with two backbones, while remaining competitive across nine additional mathematical, systems, and heuristic-design tasks. Code will be released at https://github.com/BoYuanVisionary/LabBook.