Science & Mathematics

Coding agents can invest substantial effort in repository exploration and testing yet still produce unsuccessful patches. Existing online steering methods review trajectories periodically or invoke an advisor when a monitor detects behavioral problems; frequent reviews are costly, and the absence of anomalous behavior does not mean that an agent is on the right track. We propose PAUSE, a framework for steering coding agents with a single planning intervention. Agents reach their first source-code edit at a median of 60% of trajectory progress, after exploration, bug reproduction, and testing. PAUSE intercepts this edit request before it executes, which requires no learned monitor, no search over checkpoints, and no extra model call (search-free). A separate Planner reads the task and the pre-edit transcript and writes one repair plan; the same agent, the Executor, resumes with this plan, keeping its context and workspace. PAUSE applies online, during a live run, and offline, from a recorded trajectory prefix. We evaluate PAUSE on SWE-bench Verified and SWE-bench Multilingual with two Executors and three agent harnesses. With Qwen3-coder as Executor and GPT-5.6 Sol as Planner, online steering improves resolution by 16.8% on Verified in OpenHands and by 14.7% on Multilingual in SWE-agent. In SWE-agent, PAUSE outperforms LivePlan, a rule-triggered advisor, by 3.2% on Verified and 3.7% on Multilingual at similar Planner cost, and is comparable to SICA's periodic overseer at about 11-15% of its Planner cost. Offline, PAUSE recovers 35.4% of eligible failed Verified tasks, and resuming from the recorded prefix uses 33-41% fewer Executor tokens than restarting.

Coherent differentiation has been introduced a few years ago as a refinement of the (categorical semantics of the) differential lambda-calculus and of differential linear logic, compatible with the determinism of computation, and hence in categories where addition of morphisms in the hom-sets is only partially defined. In this setting differentiation is implemented by a kind of ''tangent functor'', with the difference that it maps an object to its object of summable tuples rather to a tangent bundle. In our earlier work, these tuples were pairs in the case of simple differentiation, and families indexed by integers when dealing with Taylor expansion and the Fa{à} di Bruno formula. We push this idea further, considering now tuples indexed by an arbitrary commutative partial monoid satisfying a nonnegativity condition. The monoid structure on these indices plays a crucial role in the definition of the ''differential'' calculus associated with this ''tangent functor''. In that way, and without changing the general categorical infrastructure of coherent differentiation, this theory accommodates new notions of ''differentiation'' of which we describe several instances. Among many others, such notions include various forms of finite differences as well as Taylor expansion truncated at any fixed degree, and finite Taylor expansion of a fixed finite degree with a remainder.

Reliable uncertainty estimation is important across visual recognition tasks, providing information about the confidence and reliability of model predictions. Many uncertainty estimation methods obtain predictive diversity through mechanisms that can suppress intermediate features, perturb the input, or require substantial additional computation. Stochastic activation functions, where network units randomly select between multiple nonlinearities during a forward pass, have previously been used to improve optimization behavior, mitigate limitations of fixed ReLU activations, and increase output diversity. In this work, we reframe stochastic activation switching as a lightweight mechanism for uncertainty estimation. The proposed method samples between nonlinear activation functions, inducing predictive diversity while keeping intermediate feature pathways active. This enables uncertainty estimation from a single trained model without the training cost of ensembles, task-specific uncertainty heads, or architectural redesign. We evaluate the proposed approach across classification, semantic segmentation, and out-of-distribution detection using convolutional and transformer-based architectures on diverse datasets. Across these settings, stochastic activation switching improves uncertainty quality while preserving competitive primary-task performance, showing that stochastic nonlinearities provide an effective and information-preserving source of predictive variation for visual recognition.

Instruction-data selectors choose which instruction-following records are used for fine-tuning, improving alignment quality, raising training efficiency, and filtering out harmful records. A selector's choices thus directly affect the fine-tuned model's safety, yet despite their widespread use, the security of these selectors remains under-studied. We conduct the first systematic study of adaptive supplier-side attacks against utility-oriented instruction-data selectors and introduce utility-hijacking. Utility-hijacking shapes harmful records to remain compatible with a selector while preserving their malicious payload, causing top-k selection to concentrate rather than filter the poison. Achieving both is non-trivial, because optimizing a record for score tends to wash out its payload, while bluntly preserving the payload makes the record look low-value. Our generate-and-rerank procedure generates, per instruction, many payload-bearing candidates and submits only the highest-scoring survivor. Evaluated across three payloads (over-refusal, false-fact injection, and insecure code), six selectors, and victim models from Qwen2.5-7B to Qwen3-32B, utility-hijacking drives refusal from 0% to 94% and insecure code up to 56/90, while false-fact enrichment intensifies with scoring-model scale from 2.84x to 7.04x. To defend against this attack, we propose CONCERT, a targeted defense needing no content detector or poison labels. It flags suppliers selected far more often than chance allows, then re-selects under the same budget, keeping only records that two dissimilar selectors both rank highly. On a Qwen3-32B victim, CONCERT reverses the poison concentration from 6.67x to 0x (code) and 3.59x to 0.67x (facts), and reduces downstream harm by 100% and 75% respectively, at unchanged capability.

An $n$-player game in normal form is said to be of format $\boldsymbol{d} = (d_1, \dots, d_n)$ if the $i$-th player has $d_i$ options for $1 \le i \le n$, and is said to be balanced if $d_n - 1 \le \sum_{i=1}^{n-1}(d_i - 1)$. The Nash equilibria of the game correspond to solutions of the system of indifference equations for the game, with payoff differences as coefficients and strategy probabilities as variables. As such, the mixed discriminant of this system gives some information about which games could have an unusual number of totally mixed Nash equilibria. This mixed discriminant can be regarded as an $\mathcal{A}$-discriminant for some matrix $\mathcal{A}$ in the sense of [Gelfand--Kapranov--Zelevinsky 1994], and the convex hull of the columns of $\mathcal{A}$ is a polytope, which we call the Nashtope. We study this $\mathcal{A}$-discriminant for each balanced game format with at least three players. This discriminant is a subset of the Nash discriminant variety introduced in [Abo--Portakal--Sodomaco 2026]. We compute the codimension and degree of this discriminant for all balanced game formats, including on-boundary formats, using the Euler obstruction formula in [Matsui--Takeuchi 2011]. Our $\mathcal{A}$-discriminant locus corresponds to a special multigraded Hurwitz locus studied in [Pratt--Sodomaco--Sturmfels 2026]. As a consequence of [Pokidkin 2026], we obtain the codimensions and degrees of the irreducible components of the Nash discriminant variety for each balanced game format with at least three players, fully resolving Problem 3.8 of [Abo--Portakal--Sodomaco 2026].

Over the past decade, substantial efforts have been devoted to developing machine learning (ML) methods for extracting materials data and, where possible, materials knowledge directly from the scientific literature. This review traces the evolution of these methods from heuristic, rule-based pipelines through natural language processing (NLP) models and transformer-based encoders to, lastly, large language models (LLMs) and agentic systems capable of extracting materials information across multiple modalities, including text, tables, and figures. Beside construction of large, queryable materials databases and knowledge graphs that support downstream materials discovery, we discuss how these advances have enabled new capabilities, such as, generic question answering across materials science, AI co-scientists and chatbots for research assistance, and closed-loop optimization and discovery when integrated with autonomous laboratories. We critically assess the datasets generated using these approaches, their extraction accuracy and computational costs, and the respective strengths and limitations of different methodologies, and even provide practical guidelines for developing more reliable and scalable materials data extraction pipelines. Finally, we identify key open challenges, including the extraction of full materials structure-property-processing-performance tetrahedron at the sample level, the scarcity of reliable benchmark datasets for systematic evaluation, and the resolution of non-standard materials terminology across different document modalities.

Existing social bias benchmarks evaluate large language models mostly on fluent, well-formed English, reflecting native speaker usage and leaving non-native writers unrepresented. To address this gap, we investigate how grammatical errors by non-native writers affect social bias. We develop a controlled framework that injects common learner errors into prompts and apply it to three bias benchmarks covering stereotype alignment and demographic preference across gender, race, and religion. We find that grammatical errors increase bias, and the effect grows with the error rate. Among the four common error types we consider, preposition errors produce the largest and most consistent increase in stereotype alignment. Our analysis suggests that errors increase bias when they are more likely to interfere with comprehension. Finally, fine-tuning on erroneous texts reduces the bias increase caused by grammatical errors. Our findings show that bias evaluations based exclusively on well-formed prompts may underestimate the bias experienced by users whose English contains grammatical errors.

Haussler's covering lemma (JCTA 1995) gives the sharp $(C/δ)^d$ metric-entropy bound for binary VC classes under normalized $\ell_1$ distance. We prove a sharp global dynamic analogue for bounded real-valued classes in normalized $\ell_p$, for every $1\le p

Automated radiology report generation for multi-sequence brain MRI holds strong promise for alleviating radiologists' workload, yet it remains a demanding problem: brain MRI studies are inherently volumetric, each pulse sequence carries complementary diagnostic information, and annotated data are critically scarce for low-resource clinical languages. Existing brain MRI-report resources are still dominated by English reports, translated clinical text, or datasets adapted from task-specific imaging benchmarks, leaving routine low-resource clinical reporting and heterogeneous multi-sequence MRI protocols insufficiently represented. To bridge this gap, we introduce ViBMR, to the best of our knowledge, the first large-scale multi-sequence brain MRI dataset paired with routine Vietnamese clinical reports, comprising 3,649 patients and 11,875 MRI series across T1, T2, FLAIR, and contrast-enhanced T1 sequences. Building on ViBMR, we establish a comprehensive evaluation setup for brain MRI report generation, covering representative medical vision-language models, multi-sequence inputs, and feature aggregation strategies. The results provide key insights into dataset characteristics and model behavior, while demonstrating the necessity of native low-resource clinical datasets for reliable Vietnamese-origin brain MRI report generation. Our dataset and baseline models will be released for non-commercial research.

LLMs can act as scientific agents using tools to consult specialists and run experiments. Yet specialist information must enter their context, generally as text, images, audio or video: a bottleneck between models with rich, high-dimensional representations. We advocate direct representation exchange: a mindmeld conveying many properties, including information an LLM would not think to request. It should be extensible to new specialists, flexible across questions and tasks, and compositional across independently connected specialists without joint training. The Latent Interconnect Protocol (LIP) learns connectors mapping specialist representations into LLM internal states. Training only connectors keeps this practical as LLMs grow and evolve. We connect three protein and three small molecule specialists to Gemma 3 (4B and 12B parameters) and Kimi K2.6 (1T total, 32B active), freezing all pretrained models. Independently trained connectors provide useful information without changing the LLM or existing connectors, supporting familiar and new question types about unseen objects. We demonstrate zero-shot transfer to cross-domain tasks without joint training. Task performance and compositionality generally improve with LLM capability, with especially strong gains in zero-shot task transfer. Limitations remain: current LLMs were not trained for specialist representations as native inputs; access does not ensure reliable interpretation, integration or use in reasoning. We call on specialist developers to provide consistent representation interfaces and, especially, on LLM developers to train models to receive, combine and reason with these inputs. These efforts could expand scientific agents' information while preserving conventional tools' flexibility. Data, pretrained connectors, and code for training, evaluation and interactive inference: https://github.com/PolymathicAI/AI-Mindmeld

Code-switching is a natural linguistic phenomenon, yet its impact on high-level reasoning tasks remains largely unexplored in under-resourced languages like Romanian. Natural language processing tasks, such as satire detection, rely heavily on nuanced cultural and syntactic cues that large language models struggle to capture. In this paper, we investigate four methods (i.e., random replacement, phrase translation, dependency parsing, and large language models) for generating synthetic code-mixed Romanian sentences with English, German, French, and Italian. We start with an existing dataset for satire detection, compiled from Romanian news articles and annotated at the sentence level by humans, and augment it with over 600k samples, namely CS-SeLeRoSa. We provide an in-depth analysis of the dataset generation process. Then, we evaluate several multilingual large language models across zero-shot, chain-of-thought, and fine-tuning regimes on both the synthetic corpus and a human-adjudicated gold benchmark, namely CS-SeLeRoSa-Gold. While older open models suffer substantial degradation (dropping to 37%-55% macro-F1), newer models like Gemma 4 31B achieve 68.1% zero-shot, rivaling frontier models like GPT-5 mini (69.5%). Furthermore, fine-tuning significantly reduces formatting errors and improves the results, increasing Gemma 4 31B to 77.5% on CS-SeLeRoSa and 72.6% on the gold standard, outperforming frontier zero-shot baselines.

Motivated by the recent progress in correlated Chern insulator physics in moiré systems, we investigate the half-filled triangular-lattice Hofstadter--Hubbard model and its electron fractionalization. Among various competing non-magnetic states, increasing the orbital flux stabilizes and broadens the $\mathrm{U}(1)_2$ chiral spin liquid phase in the Mott regime. In this exotic regime, the electron is fractionalized into gapped spinons and gapped chargons, whose effective bands are linked to the orbital flux. The convolution of the narrow spinon bands with the gapped chargon excitations produces rich and discernible structures in the electron spectral function and the density of states. These results not only identify the orbital flux as a control parameter for the detailed patterns of electron fractionalization and the nature of the Mott transition, but also provide the key spectral characteristics for the angle-resolved photoemission and scanning tunneling studies of the relevant systems, linking the free-particle band topology to intrinsic topological order in the correlated regime.

As LLMs are increasingly deployed in value-sensitive contexts, aligning them with human values has emerged as a post-training objective in NLP, aimed at instilling additional values and guiding model behaviour. However, current evaluations remain narrow. Because values are interconnected, it is unclear how fine-tuning for specific values alters unrelated capabilities and behaviours. In this work, we adapt four instruction-tuned LLMs across individual Hofstede's value dimensions and evaluate them on both cross-value transfer and broader behaviours. We find that value-specific training produces unintended side effects on untrained values, risk preferences, social cognition, and safety traits. Consequently, we suggest that value alignment after post-training should incorporate broader behavioral evaluations, and we highlight the need to investigate where additional value training might best fit into the pipeline.

Nuclear power is increasingly seen as necessary for sustainable development, but this role depends on informed public engagement and support. Building this support requires understanding what citizens want to know, as reflected in their spontaneous online searches. To map this demand, we present the first long-term (2004-2024), prefecture-level analysis of Google Trends data on nuclear power in Japan. The results indicate that in routine periods, search interest followed a host-locality pattern, being significantly higher in prefectures hosting nuclear plants, with little spillover to their neighbors. Immediately after the Fukushima Daiichi accident in March 2011, this pattern changed, and interest instead declined with distance from Fukushima, before the host-locality pattern returned from 2012. During March 2011, searches leaned heavily toward general and live information rather than guidance on what to do. Their framing also varied with proximity, as those near Fukushima prioritized live updates and formal countermeasures, while others leaned toward background information and personal-scale questions. These patterns and the shift between them provide a behavioral evidence base for designing communication strategies in Japan and countries weighing the use of nuclear power. More broadly, the finding that searches vary with proximity can inform public communication in other geographically anchored crises.

As large language models (LLMs) continue to grow in scope and capability, privacy concerns are on the rise. Since users increasingly interact with a single general-purpose system across many distinct contexts, there is a natural tendency to pursue a one-size-fits-all solution to privacy. Unfortunately, this overlooks a crucial insight from decades of privacy research: privacy is inherently contextual. In this paper, we argue that existing work in prompt privacy is drifting towards attempts to design a unified framework, which is highly misleading and can foster a false sense of security. Furthermore, we argue that even future work in this direction needs to wrestle with fundamental tensions across various desirable properties of prompt privacy. Thus, such continued attempts towards a unified framework need to be reconsidered. We recommend three critical paths forward: application-specific deployment of LLMs, context-specific evaluation of privacy frameworks, and better privacy literacy.