Papers

Semantic uncertainty quantification for large language models rests on a common template: sample several answers, measure how much they agree, and treat disagreement as uncertainty. We first formalize this template as two separate roles: an operator that compares two answers, and an aggregator that combines all pairwise comparisons into a scalar. Existing methods differ almost entirely in how they aggregate, while taking the operator off the shelf, typically an NLI model or a generic sentence encoder. We show that this reliance on off-the-shelf operators is the primary bottleneck of semantic UQ: they do not accurately measure factual equivalence of multiple answers to the same question. We resolve this with a deliberately simple recipe: a single encoder trained contrastively to isolate the targeted fact, utilizing synthetic data generated by an LLM and dataset both disjoint from all evaluation settings. Integrating the resulting operator into existing methods improves performance on 120 of 126 evaluation settings (95%) spanning 18 model dataset combinations across language and vision-language models. The best variant reaches 0.76 mean AUROC against 0.68 for the strongest baseline, while replacing the quadratic cross-encoder comparisons of entailment-based operators with one encoder pass per answer. The uniformity of the improvement supports the view that the operator, not the aggregator, is the limiting factor. The same operator also improves single generation token-level estimators: the norm it assigns to each token measures how much that token bears on the answer, and reweighting token log-likelihoods accordingly sharpens the estimate.

Greenhouse climate control balances economic return with maintaining temperature, humidity and CO2 within crop-adapted growth ranges. Conventional reinforcement learning (RL) greenhouse controllers use fixed reward penalties to limit climate constraint violations, yet such heuristic penalties cannot explicitly constrain long-term cumulative violations. Poorly tuned weights either lead to overly conservative policies and lower yields, or fail to suppress persistent climate deviations that harm photosynthesis and induce crop diseases. To address this issue, we formulate greenhouse climate regulation as a Constrained Markov Decision Process (CMDP) and use a Lagrangian safe RL framework RCPO-PPO to separate economic optimization and cumulative safety constraints, enabling adaptive penalty adjustment without manual tuning. To handle strong nonlinear, time-varying coupling between greenhouse microclimate and crop growth, Kolmogorov-Arnold Networks (KANs) replace Multi-Layer Perceptrons (MLPs) as policy and value approximators for improved nonlinear representation. Sinusoidal cyclic time features are embedded in observations to capture diurnal environmental periodicity. Simulations use a classic winter lettuce greenhouse model driven by 40-day real weather disturbances. Compared with vanilla penalty-based PPO, our method cuts cumulative climate violations by 18.65% and raises lettuce economic profit by 2.91%, keeping violations stable near the safety threshold. This decoupled CMDP optimization with KAN-based policy representation mitigates long-term climate risks and boosts planting profits, offering a constraint-aware control strategy for precision greenhouse cultivation.

Many physiological time series, such as cardiac and brain recordings, exhibit cyclostationarity: their statistics vary periodically with an underlying cycle phase. Corruption from motion, poor contact, and physiological interference obscures morphology needed for diagnosis, making signal restoration essential. Existing diffusion approaches condition on corrupted observations alone and must learn cyclic structure implicitly. We instead propose two inductive biases which encode cyclostationarity: a shift-covariant wavelet representation and dense per-sample phase conditioning inferred from the corrupted input. We further introduce a training-free cyclostationarity index that quantifies phase structure and predicts when phase conditioning will help. Finally, we propose antithetic coupling of reverse trajectories to reduce sampling variance while achieving comparable performance with fivefold fewer network evaluations. Across modalities, our results show that explicitly encoding measurable cyclic structure improves physiological time-series restoration.

Large language models can generate functionally correct code that still contains security weaknesses, motivating repair pipelines that first diagnose a weakness type before deciding how to fix it. The Common Weakness Enumeration (CWE) provides a standardized vocabulary for such diagnoses, but asking an autoregressive language model to generate a CWE label and extracting it from the response raises questions about output validity, speed, and cost, as well as accuracy. We evaluate Jev, a decision model that instead selects directly from a declared set of candidates and returns a probability for each, against six open-weight autoregressive models and a frontier proprietary model, GPT-5.6-Sol, on a controlled 50-way CWE classification task over 1,916 CyberSecEval benchmark examples. Jev outperforms all six open-weight baselines on every classification and ranking metric, while its comparison with GPT-5.6-Sol depends on the metric: GPT-5.6-Sol achieves higher Top-1 accuracy and Macro-F1, whereas Jev achieves higher Top-3 and Top-5 accuracy and a nearly identical MRR, at $6.27\times$ lower median API latency and $55.9\times$ lower estimated API cost. We further build JevVibe, a diagnosis-guided repair agent that uses predicted CWE labels to repair code generated by Qwen2.5-Coder-32B-Instruct. With Jev providing the diagnosis, the agent increases the detector-measured security pass rate from 63.5% before repair to 70.7%, compared with 66.1% for LLM-guided repair. These results show that JevVibe is effective at improving the security of generated code, with Jev providing reliable and efficient CWE classification.

Estimating mutual information (MI) from samples is a central objective in a variety of scientific fields. Modern neural estimators are accurate in the large-data regime, but they fall short when data is scarce, and each must be fit anew for every distribution under study. Current estimators are moreover tied to specific data types. These constraints limit their adoption in many applications where per-distribution training is impractical and sample sizes are small. We present ALICE, a foundation model that removes per-distribution training, while achieving competitive estimation accuracy. Trained exclusively on a broad family of synthetic distributions, ALICE acts as an in-context estimator of rectified-flow velocity fields: conditioned on samples of an unseen distribution, it estimates that distribution's velocity field without any explicit training. MI is then obtained through a fixed identity that integrates the squared difference between the joint and conditional fields. We validate ALICE on a standard, challenging benchmark and apply it in three domains, biology, genetics, and neuroscience, whose data the model has never seen. For the first time, we show that a single model closes the gap with neural estimators trained separately for each distribution, while natively supporting different data dimensionality and sample cardinality, enabling zero-shot MI analysis across scientific domains.

Formalizing research-level stochastic optimization in Lean requires both an algorithm model and domain theory connecting foundational libraries to convergence proofs. Revising a model to restore provability can change the mathematical claim. We introduce ProofLoom, a fully automated LLM-agent system for Proof-Obligation-Driven Theory Construction. Given a published algorithm, target theorem, and source proof, ProofLoom autonomously constructs the Lean model and supporting theory. Open proof obligations drive the development of definitions, interfaces, lemmas, and proof plans. Signature contracts record evidence and obligations for model revisions; an independent Judge rejects unsupported assumptions and weakened conclusions. Planner expands the published argument into intermediate claims, and Audit checks whether the Lean proof follows it. Across tasks, SOptLib accumulates verified mathematics and construction experience: reusable results are extracted, generalized, and verified, while modeling decisions and failed proof routes are recorded. Later tasks retrieve these results and records and contribute new developments, forming a cycle of construction, accumulation, and reuse. On fifteen textbook and research-paper tasks, ProofLoom obtains mean human ratings of 6.3/7 and 6.4/7, compared with 4.9/7 and 5.0/7 for the strongest of six baselines. Across 33 developments, it produces 490,693 lines of algorithm-local Lean code with no sorry. The formalizations also expose 28 incorrect formulas, proof gaps, and algorithm-analysis mismatches in published sources across 22 developments, each with checked evidence.

Large-scale learning programmes generate records that make learner participation observable across different activities. Participation points are commonly used to record and encourage such participation, but their value may extend beyond the activities for which points are awarded. Existing evaluations often examine gamification outcomes within the activities or learning environments in which the game elements are implemented, providing limited evidence about whether early participation points contain information about later participation outside the points system. This study examines whether early participation points can provide information about learners' later participation in a self-paced learning track that does not award participation points. Using anonymised records from 876 learners in a large-scale remote software-upskilling internship, we examined participation points generated from live-session attendance and poll responses against later self-paced course completion. The primary analysis used the 438 learners who earned at least one point during the first week, while the full cohort was retained for the no-point analysis. Week-one participation points distinguished learners who later completed a self-paced course with an AUC of 0.89, increasing to 0.95 by the fourth week. Similar AUCs were observed at both stages of the self-paced course sequence, while the absence of week-one points identified learners who did not start or did not complete a self-paced course with 95% precision. These findings indicate that early participation points can provide information about later participation outside the activities that generate the points. Such information can help large-scale learning programmes identify learners who may require timely attention while learning is still in progress, without treating participation points as a measure of overall learner engagement.

We study the consequences of membership comparators that exclude one possible membership vector, deterministically or with a relative advantage over uniform guessing. For every polynomially bounded arity, a randomized polynomial-time comparator of error at most $(1-1/poly(n))2^{-t}$ gives $ NP/ poly\cap coNP/ poly$ recognition with common advice. The proof uses limited independence, polynomial occurrence certificates, and a self-contained positive-relation advice transfer. For SAT at arity $O((\log n)^d)$, both this relative-gap hypothesis and deterministic comparability imply $PH=S^{NP}$, the uniform bound $PH\subseteq BPTIME(2^{O((\log n)^{d^2})})$, and symmetric verification with polynomial-length certificates and an oracle-free deterministic $2^{O((\log n)^d)}$ predicate. Polynomial-advice deterministic decoding has the same exponent $d$. Applying the randomized simulation to an unconditional diagonal language yields, for every fixed $\varepsilon>0$, $\mathrm{BPP}\subsetneq BPTIME(2^{O((\log n)^{d^2+\varepsilon})})$, without advice. The larger clock remains subexponential under every fixed number of self-compositions. A layered oracle satisfies deterministic comparability and $NP^O=coNP^O$ but excludes randomized NP algorithms with smaller logarithmic power, establishing a relativized limit on the SAT exponent $d$. This expanded version also develops the full weak-advantage regime, where saving $2^{-O((log n)^d)}$ gives randomized SAT exponent $d$ and PH exponent $d^k$ at fixed level $k$; the quasipolynomial and exponential hierarchy consequences; binary-comparator advice bounds; and the certificate-length boundary between the randomized regimes. Under deterministic comparability, uniform deterministic promise-unique search additionally gives $UEXP=EXP$. The ordinary second-level collapse $PH=\Sigma_2^p$ for $d>1$ remains unproved.

In 2023, AI models answered from training data and hallucinated when it ran out, and businesses were told to seed that knowledge. Models' training knowledge has since given way to live web search, and the advice followed it there: answer-engine optimization, or AEO, now tells businesses to scatter breadcrumbs across forum threads, listicles, and off-site citations, so AI engines are likelier to surface and recommend them. But being surfaced is no longer enough: an agent opens the results and reads them before deciding, and one buyer question sends it through several rounds of search and fetch. What decides the outcome at this drill-down step is whether the agent can fetch and read the business's own site: agent experience (AX). We argue that AX is the new AEO. We run 37,927 agent journeys, each a buyer question about a business, across four independent harnesses over 1,056 real businesses, matched on fame, prior model knowledge, and two AEO proxies, then split based on their AX level. Only 7-10% of the finished answer comes from the model's training knowledge, whether or not the site is readable. Agent-ready businesses have answers built from their own pages 78% of the time against 56% and are clearly recommended 1.9x more often, while a grounded answer about a not-agent-ready business costs the agent 64% more on average. Holding business, harness, and question fixed, answers built from the site are 41% more accurate on average. The dominant failure is not fabrication but omission: web-built answers are 3.7x more likely to contain none of the facts the buyer asked for. Baselines differ sharply across the four harnesses, with clear-recommendation rates varying sevenfold from stack to stack, yet the recommendation gap holds in every one. In the agentic web era, being readable beats being talked about, and improving a site's AX is the strongest lever a business has.

AI-powered lifelogging wearables are emerging as a new class of consumer devices that transform everyday experience into searchable, AI-curated memory archives. We study early public sensemaking around these systems at the moment of their market entry, using the Looki L1 as an empirical lens. Analysing large-scale Chinese-language and English-language social media discourse (N = 5,053 comments), we combine topic clustering with inductive thematic analysis to examine how users interpret the social, moral, and political implications of AI-mediated memory. Across contexts, users reference dystopian surveillance imaginaries, express privacy resignation and bystander concerns, and debate assistive value alongside consumer logics. English-language comments more often framed these devices through interpersonal power, evidentiary use, and hacking anxieties, while Chinese-language comments more often foregrounded labour exploitation, governance surveillance, and technological inevitability.

Few-shot visual anomaly detection is fundamentally a visual comparison task, requiring fine-grained inspection of a query against normal references. Many recent methods based on large vision-language models (LVLMs) emphasize comparative reasoning through language chain-of-thought. Yet discrete, abstract descriptions may underrepresent dense, fine-grained visual differences, leaving a gap between visual comparison and its expression in language. To address this gap, we propose Visual Difference DeepStack (VD-DeepStack), which explicitly conditions language reasoning on query-reference visual differences. Specifically, we fuse DINO features with the LVLM visual hierarchy to strengthen fine-grained representations, then construct dense difference evidence from residuals between query features and softly matched reference features. The difference-evidence path injects spatially weighted difference vectors into query-image states at multiple decoder depths, while an auxiliary visual-context path provides fine-grained appearance information to support their interpretation. Experiments on 4 industrial and 2 medical anomaly benchmarks demonstrate substantial improvements in few-shot anomaly detection over baselines relying on textual comparative reasoning. These results support mitigating the visual comparison-reasoning gap through the joint design of comparison representations and their integration into the decoder. Code will be released upon acceptance.

Proactive dialogue requires agents to continually adapt their policies to user feedback while progressing toward task objectives over multiple turns. To move beyond imitation learning on static datasets, recent approaches use user simulators to collect interactive data for policy optimization. However, many simulators do not explicitly model the evolution of user cognition, limiting the consistency and state dependence of feedback across turns. Moreover, representing each action only by a high-level strategy label overlooks the large utterance space and cannot distinguish alternative realizations of the same strategy. To this end, we jointly design a $\textbf{Cog}$nitive User $\textbf{Sim}$ulator $\textbf{(Cog-Sim)}$ and $\textbf{C}$ognitive-$\textbf{S}$tate $\textbf{T}$ransition–Driven $\textbf{P}$olicy $\textbf{O}$ptimization $\textbf{(CSTPO)}$. Cog-Sim maintains the user's cognitive and affective states and generates responses through constrained state transitions across turns, so feedback depends on both the realized utterance and the user's current state. CSTPO organizes each action as a hierarchical strategy–utterance representation: a high-level strategy label constrains utterance sampling, and utterances are optimized within each label. Sparse complete-branch sampling reuses shared dialogue prefixes and estimates separate strategy-level and utterance-level advantages, enabling fine-grained optimization at both levels. Across three tasks, Cog-Sim exhibits monotonic dose–response relationships and is preferred over prompt-based simulators for naturalness. CSTPO improves Qwen3-14B's performance to a level comparable to that of GPT-5.5-based planning methods.

Flow-based Vision-Language-Action (VLA) policies are typically trained by behavior cloning and thus do not explicitly optimize long-term task return. Critic guidance steers generation toward higher-value actions, but existing methods differentiate the critic through a one-step surrogate of the sampler and back-propagate a critic ensemble at every flow step. In contrast, here we propose Adjoint Guidance Flow (AGF), which amortizes trajectory-aware critic guidance into a lightweight guidance network while preserving the pretrained VLA policy. Specifically, we formulate critic-guided flow generation as a deterministic optimal control problem, whose optimal guidance is a costate that carries the terminal critic gradient back through the remaining flow, and regress the guidance network onto this costate while keeping both the VLA and critic frozen. This design provides favorable memory and throughput scaling during training, and inference needs one guidance-network forward pass per step, without the critic ensemble, back-propagation, or adjoint computation. Across LIBERO, RoboCasa, and LIBERO-Pro, AGF consistently improves pretrained VLAs, remains competitive with critic-guidance and policy-fine-tuning baselines, and is the most robust method when a single guidance strength is deployed across tasks. Compared with QGF, AGF runs $3.6\times$ faster per guidance step with $7.0\times$ fewer parameters, with comparable and even better performance, showing that critic guidance can be trajectory-aware and lightweight.

The inference efficiency of Multimodal Large Language Models (MLLMs) is severely constrained by massive visual token sequences induced by high-resolution inputs, with computational cost scaling quadratically. Existing approaches primarily focus on downstream token compression, while overlooking a fundamental upstream inefficiency: input resolution is treated as a static, task-agnostic hyperparameter. We propose Task-Conditioned Resolution Routing (TCRR), which formulates visual compression as a task-conditioned decision and employs a lightweight cross-modal router that conditions backbone visual representations on textual semantics via feature-wise modulation and cross-attention to predict the minimal sufficient compression level per query. To support this, we curate a dataset of 500k samples across 12 task categories, labeled via a teacher-oracle pipeline to approximate Pareto-optimal compression scales. Extensive experiments across diverse architectures show that TCRR achieves a superior efficiency frontier, specifically reducing visual FLOPs by 40.9% and latency by 53.7% on Qwen3-VL-8B while preserving competitive performance. Further analysis of scaling behavior confirms that dynamically routing visual compression enables optimal resource allocation without modifying the MLLM backbone.

Future exogenous variables provide valuable information for forecasting endogenous time series. Existing covariate-aware methods primarily learn the direct influence of exogenous variables on endogenous variables. However, these effects can be complex and change with the pattern of the exogenous variables, making them difficult to capture. Beyond this perspective, we observe that a given exogenous pattern often co-occurs with only a small set of endogenous response patterns. These associations motivate a strategy that matches future and historical exogenous patterns and uses the corresponding endogenous patterns to enhance forecasting. However, in real-world forecasting scenarios with multiple exogenous variables, each exogenous variable provides a distinct dimension for matching, creating a dilemma for this strategy between precise matching and sufficient historical support. To bridge this gap, we propose XMatch (EXogenous MATCHing), a covariate-aware forecasting model that realizes the aforementioned strategy through a tree-structured matching process that adaptively adjusts the number of exogenous variables used as matching conditions. Specifically, we first introduce the ProtoTree Creator, which organizes historical correspondences between exogenous and endogenous patterns into a ProtoTree, whose deeper levels incorporate additional exogenous variables for matching. For forecasting, we then design the ProtoTree Matcher, which uses future exogenous variables to query the ProtoTree and adaptively determines how many exogenous variables to use for matching based on exogenous pattern similarity and historical support. Finally, the matched endogenous patterns are used as explicit historical evidence to enhance forecasting. Extensive experiments on 12 real-world datasets demonstrate that XMatch outperforms state-of-the-art baselines.

Exponential stabilization is investigated for a hetero-dimensional parabolic system in which a one-dimensional shoreline subsystem, subject to a constant input delay, is bidirectionally coupled with unactuated two-dimensional bathymetric dynamics. A spectral decomposition separates the finite-dimensional critical shoreline modes from the exponentially stable residual dynamics. State feedback is designed on the critical subspace and lifted to the full shoreline space. An auxiliary transport equation represents the delay, while a predictor removes the explicit input delay from the nominal predictor dynamics. Exponential stability of the full coupled system is established by combining shoreline stabilization and bathymetric coercivity using weighted input–output estimates and a small-gain condition on the hetero-dimensional interaction. Numerical simulations show that uncompensated delayed feedback can strongly amplify both shoreline and bathymetric responses, whereas predictor compensation restores decay of the critical shoreline dynamics and prevents delay-induced amplification from propagating to the unactuated bathymetric subsystem.

It is unclear how Neural Language Models (NLMs) acquire the structural meaning encoded by grammatical structures that is independent of lexical semantics. We propose a statistical learning process in which learned dependency structures themselves become new distributional units for subsequent statistical learning. Under this account, once a dependency structure is acquired, the model tracks its contextual distributions. These contextual features reflect the semantic properties of a composite structure. To test this hypothesis, we design a synthetic grammar in which each grammatical structure has distinct contextual distributions that cannot be recovered from the distributional statistics of their component tokens alone. We train a series of BERT-style masked language models on this grammar and examine their developmental trajectory. The results show that models can successfully learn the contextual distributions of composite dependency structures even though they cannot be inferred from token statistics alone. Developmental analysis further reveals a clear developmental trajectory. The learning of the dependency relations that define a grammatical structure consistently precedes the learning of its contextual features. These findings suggest that statistical learning in NLMs is not merely the accumulation of token co-occurrence statistics, but a process in which learned dependency structures become new units of distributional learning. We argue that this process provides a statistical-learning account of how NLMs solve the compositionality problem in language. Finally, we discuss the possibility that this statistical learning process provides an explanatory theory on how language cognition could emerge from pure distributional statistics.

Recent 3D generation models can produce accurate geometries while still struggling to reconstruct detailed textures. We propose a diffusion-based native 3D material generation model TaoTex, which faithfully recovers intricate textures through tailored strategies and improvements. First, we develop a data construction agent to create high-frequency textured 3D assets to bridge the data gap in public datasets. Training with these data significantly enhances the ability of TaoTex to recover challenging details such as text and patterns. Second, we design a multi-level feature fusion (MLFF) module to adaptively integrate local and global features of the conditional input, providing more complete texture cues for the diffusion model and thereby enhancing reconstruction fidelity. To alleviate VAE reconstruction errors, we adopt a latent-to-pixel space loss transition, further improving the pixel-level details and generation quality. Finally, we scale TaoTex to multi-view inputs by incorporating learnable viewpoint embeddings, achieving accurate and consistent material reconstruction across views. Extensive experiments demonstrate that our method significantly outperforms existing approaches in preserving texture details in both single- and multi-view settings.

Memorization has been proposed as a mechanism to explain how language models fit the tail of their training distributions, but its training dynamics are not understood well. In this work, we take a fine-grained look at memorization by decomposing the loss trajectory of memorized sequences over training and model parameters. Across the Pythia family, we study memorization of duplicated training sequences (recitation) and rare ones (recollection). We find that memorization in both cases is characterized by sequence-level gradient alignment, though recitation suffers from misalignment with other training influences which causes forgetting, explaining the necessity for higher duplication of these examples. We further show that the lower model layers are the most involved in memorization and forgetting. Predicting memorization, our decomposition improves over a cross-entropy baseline, especially in larger models and early in training. Intervening on a small set of highly influential parameters we are able to ablate memorization in the final model. Together, these findings advance our understanding of how memorization develops during training and offer insights for predicting and intervening on it.

Machine learning has made strong progress on music tasks, both as assistive tools and as creative partners. However, most systems train on multitrack corpora that emphasize pop and rock. Jazz, with improvisation at the core of its practice, still lacks a well-annotated corpus of clean per-stem combo recordings on standards. We introduce JazzSAMBA (Jazz Synchronous and Asynchronous Multi-take Band Audio) to fill this gap: the first originally recorded jazz-combo multitrack dataset of standards with asynchronous (overdubbed) and synchronous (live ensemble) protocols, preferred and alternate takes chosen by the musicians, and timed annotations for bars, chords, sections, and soloists. JazzSAMBA covers 76 standards by eight musicians on drums, bass, piano, trumpet, and saxophone, with per-stem audio, mixtures, and MIDI. It can support chart-conditioned accompaniment, combo source separation, and form-aware music information retrieval. We demonstrate the dataset on two tasks: a jazz combo source-separation baseline and a chart-conditioned accompaniment ablation. The dataset, code, and samples are linked from the project demo page.

Large language model (LLM) agents are evolving from tool-calling systems that execute isolated instructions into task-oriented agents that pursue user goals through sustained, multi-step interactions. However, existing benchmarks for personalized tool use largely assess isolated calls or reactive execution, leaving unclear whether agents can formulate, execute, and revise an explicit plan while preserving user preferences throughout long-term interaction. To address this gap, we introduce PDEU-Bench (Personalized plan Definition, plan Execution, and plan Update Benchmark), a benchmark for evaluating the complete planning lifecycle of personalized tool-using agents. PDEU-Bench comprises 214 long-horizon interaction tasks spanning 12 everyday domains and 94 tools, with stage-specific assessments of preference adherence and plan quality. Extensive evaluations of 15 representative open-source and closed-source LLMs reveal a pronounced gap between local tool execution and dynamic planning: LLMs can often instantiate preferences in individual calls, yet struggle to construct coherent plan definition and plan update. We further evaluate mainstream personalization and memory-augmentation methods. Although these methods improve particular stages, none of the evaluated methods reliably propagates user preferences throughout the complete lifecycle, and their gains frequently fail to transfer to subsequent execution. Fine-grained error analysis further reveals that preference omissions and conflicts persist throughout the planning lifecycle, highlighting the need for future research to parameterize LLMs with preference-aware information retrieval and memory capabilities. We provide the relevant code and data in the appendix to support future research.

Language models are increasingly used to sample from a specified distribution, for instance, to simulate survey respondents or generate synthetic data. Instruction-tuned models can state such a distribution correctly and still fail to sample from it. Prompting and changes to decoding reduce this mismatch only partly, which motivates training with policy optimization. Group relative policy optimization (GRPO) is a natural fit for this problem because it already samples a group of rollouts per prompt, and the group's empirical distribution can be compared with the target. However, scoring the group as a whole gives every rollout the same reward. Group-relative centering then sets all advantages to zero, and the model receives no learning signal. To give each rollout its own signal, we introduce the witness advantage, a per-rollout advantage derived from maximum mean discrepancy (MMD). It trains a model to match a target distribution over a finite set of outcomes. The MMD between the model's distribution and the target has a witness function that measures how over- or under-produced each outcome is. Each rollout's advantage estimates the negative witness at its outcome, so a rollout is rewarded for an outcome the group under-produces and penalized for one it over-produces. The witness advantage is computed in closed form from the group's outcome counts, and we use it as the reward in GRPO. On unseen target distributions, training with the witness advantage substantially reduces the total variation distance to the target while largely preserving the model's general capabilities.

Structural analysis is a core method in Petri Net (PN) research, offering a perspective complementary to state-space techniques while avoiding many of their limitations. It is well studied for classical PNs but far less for High-Level Petri Nets (HLPN). Symmetric Nets (SN), a common HLPN formalism, use compact annotations to encode behavioral symmetries, enabling the construction of a symbolic reachability graph (and a lumped Markov chain in stochastic SN) and the execution of symbolic discrete-event simulations. During the past two decades, structural techniques tailored to SN have been developed, notably supported by the SNexpression tool. This tool implements a formal calculus designed for the computation of symbolic structural relations, including, but not limited to, conflict relations and causal dependencies. Here, we focus on using this calculus to verify semi-automatically symbolic structural invariants, a task currently feasible only for certain restricted SN subclasses. We focus specifically on (semi)flows and briefly discuss an approach through which a flow generative family can be generated, at least theoretically. We further briefly outline a framework for the formal verification of a broader class of invariant properties. An extended formulation of the SN formalism is employed, which demonstrably satisfies the closure property with respect to fundamental functional operators. The core concepts are elucidated by means of representative examples throughout the exposition.

We prove that the state space of every well-formed session type, quotiented by strongly connected components, forms a bounded lattice; n-ary parallel composition yields product lattices. Two consequences follow: duality preserves the lattice up to isomorphism, and Gay-Hole width subtyping corresponds to lattice embedding for non-recursive types. We validate this on 108 benchmark protocols across networking, databases, distributed systems, AI, and fault tolerance: all form lattices, 93 distributive, 15 non-distributive. Mechanised in Lean 4 with two independently developed tool implementations.

Asynchronous communication is a fundamental feature of modern distributed systems, where messages are emitted without requiring immediate synchronization with receivers. In process calculi, this behaviour is typically modelled by separating message emission from message consumption. At the same time, reversible computation has emerged as an important paradigm for analysing concurrent systems, enabling computations to be undone while preserving causal dependencies between actions. While reversible semantics have been extensively studied for synchronous process calculi such as CCS, their integration with asynchronous communication remains largely unexplored. In this paper we investigate the interaction between asynchrony and reversibility in the setting of CCS. We first introduce CCSa, an asynchronous variant of CCS in which output actions generate explicit message entities that can later be consumed by matching input actions. We then define rCCSa, a reversible extension of CCSa obtained by adapting the framework of Phillips and Ulidowski. In rCCSa, prefixes and messages are annotated with unique keys that record message emission and consumption events, allowing computations to be reversed while preserving causal dependencies. We show that the resulting reversible semantics satisfies causal consistency, ensuring that computations can be reversed exactly up to causal equivalence. The proof relies on the axiomatic framework for reversible computation proposed by Lanese et al.

An auditor who checks whether a system's weights are frozen is checking the wrong thing. Our stationarity dichotomy says that iterative self-modification hits strict diminishing returns whenever the agent's reachable set of edits stays fixed, and can escape only if that set expands. Rewriting scaffolding (tools, verifiers, decomposition) expands what an agent reaches without touching a weight, so frozen weights buy an eventual ceiling but no stationarity along the way. The criterion also separates three regimes usually merged: search within a fixed class, test-time training that raises the ceiling itself, and scaffold rewriting between them. Audit the scaffold, not the checkpoint. The same ceiling binds sideways. Best-of-$k$ orchestration realizes the best worker's ceiling exactly: width buys rate, not budget. Re-consulting a fixed pool has a horizon computable in advance, decided by the pool alone, and the one arrangement that would beat it, a weighted vote, needs diversity real workers lack: on 30 same-family workers the failure overlap sits at its maximum, and a majority fails 23/55 (42%) of tasks. We obtain the criterion by reading refinement as gradient boosting on the residual error between draft and target, a patch or git diff, and then measuring where that reading breaks: patches compose instead of standing beside each other to be voted on, and failures overlap. What we measure is saturation. Per-round improvement decays toward zero on SWE-bench, and churn decays geometrically across 401 production sessions, a shape shared with a pre-AI human baseline that establishes the regime without identifying its cause. Both breaks are engineering choices rather than laws about code, so together they specify a harness worth building.

The Participants-as-Interfaces (PaI) methodology for system composition proposes that system participants can be regarded as interfaces. For each system in a given collection, one participant is designated to serve as its interface. When the systems are composed, these interface participants are replaced with gateways that communicate with one another by forwarding messages. We generalise the approach to partial gateways, where gateways can forward only a chosen set of messages. As for the standard PaI approach, we exploit such extended version for systems of communicating finite state machines (CFSMs). This extension is fully detailed for the binary case, and can be scaled up to the multicomposition case. We prove that many relevant communication properties (deadlock-freeness, reception-error-freeness, etc.) are preserved by PaI composition via partial gateways in case also the connection policy (i.e. the system representing the way we wish gateways interact with each other) enjoys the same properties. Such a proof turns out to be just a corollary of a property preservation result for a restricted and partial composition method, dubbed fusion-composition. Fusion-composition hence turns out to be at the heart of the PaI composition approach.

Predicting Drug-Target Interactions~(DTIs) is a central task in computational drug discovery, with direct applications in virtual screening, drug repurposing, and therapeutic candidate prioritization. Although recent deep learning methods have improved DTI prediction, many sequence-based models still process drugs and proteins independently and only combine their representations at a late prediction stage. This limits their ability to explicitly model cross-molecular dependencies between chemical substructures and protein sequence regions. In this paper, we propose a sequence-only DTI prediction architecture that combines two pre-trained language models, ChemBERTa for drug SMILES strings and ESM-2 for protein amino acid sequences, with a hierarchical interaction module. The proposed model first extracts contextual representations using pre-trained encoders, then applies 1D convolutional layers to condense local sequence patterns, followed by a sequential bidirectional cross-attention mechanism inspired by the induced-fit view of molecular recognition. Finally, attention-based pooling constructs fixed-size interaction-aware vectors for binary prediction. Experiments on BIOSNAP, Davis, and BindingDB show that the proposed model achieves the best performance on BIOSNAP, matches the best AUROC on Davis, and remains competitive on BindingDB while using only 25.2 million trainable parameters. Ablation results confirm the contribution of both the CNN and cross-attention modules, and cold-start experiments indicate promising generalization to unseen proteins and drugs.

On modern production text-to-image systems, successful policy violations are rare, and previously effective human-written seeds are often patched out. Current automated red-teamers are poorly matched to this regime in two ways: unreliable success measurement and poor exploration. First, we find that judges widely used in prior T2I red-teaming work are unreliable under vague unsafe-content targets: they either miss true violations or reward benign borderline images on hardened APIs. We therefore define strict category-specific success criteria and calibrate strong VLM judges against human labels. Second, we show that broadly used prompt-modification pipelines do not solve the exploration problem: on harder guardrail settings they remain tied to seed prompts, fail to transfer, or cannot bootstrap positive examples. We introduce RISE, which evolves reusable strategies used to generate prompts rather than rewriting them one by one. The best discovered strategies are then reused to generate attacks across new scenarios. On DALL-E 3, Nano Banana 2 (Google) and GPT-Image-2, RISE reaches up to 13% human-verified ASR; under the same calibrated evaluation, prior methods with reported ASR as high as roughly 30% fall to near zero.

For functions that are convex of higher order, classical extremalities identify Gaussian quadrature as a lower bound and Gauss–Lobatto quadrature as an upper bound for the integral. We sharpen this bracket in every odd order: if $f$ is $(2n-1)$-convex on $[-1,1]$, then the exact integral $I[f]:=\int_{-1}^{1}f(x)\,dx$ lies not merely between the $n$-point Gauss–Legendre rule $G_n[f]$ and the $(n+1)$-point Gauss–Lobatto rule $L_{n+1}[f]$, but already between $G_n[f]$ and their midpoint. This yields the certified estimator $Q_n[f]:=\frac{3}{4}G_n[f]+\frac{1}{4}L_{n+1}[f]$, with $|I[f]-Q_n[f]|\leqslant\frac{1}{4}|L_{n+1}[f]-G_n[f]|$, and the constant $\frac{1}{4}$ is sharp. We also record a complementary phenomenon in even order: for $2n$-convex functions the Gauss–Radau endpoint rules still bracket the integral, but the midpoint of the two Radau rules cannot yield a one-sided refinement. The obstruction is explained by an odd, sign-changing Radau quadrature kernel. Finally, we distinguish numerical certification from ordinary high-order approximation: we include concrete examples from spline, moment and statistical models and report experiments for truncated powers and near-pole Stieltjes kernels, where the shape certificate remains meaningful when classical smoothness-based error constants are unavailable or severely pessimistic.