Artificial Intelligence Papers

ALT Serving web queries at scale with multi-agent systems requires jointly configuring large language models, collaboration modes, agent counts, interaction rounds, and skills. However, the resulting configuration space grows combinatorially, while sparse execution rewards provide limited fine-grained supervision for interdependent planning decisions. We propose EvoRouter, a self-evolving multi-agent routing framework driven by insight-guided on-policy distillation. A dedicated expert periodically synthesizes performance-cost feedback from configuration trajectories across tasks into reusable conditional routing insights about capability requirements, returns from resource scaling, and interactions among configuration choices. A privileged teacher combines these evolving insights with current execution feedback to provide dense token-level supervision for the router's own planning trajectories through on-policy distillation. Cost-aware reinforcement learning jointly optimizes task performa

ALT Reward guidance steers a pretrained diffusion or flow model toward images or actions that a reward model prefers, at sampling time and without retraining the model for each reward. It follows the gradient of a value function, the expected reward given the noisy state. The two common estimates of this gradient, gradients through a one-step denoised estimate and learned value models, are least accurate at high noise, where the coarse structure of the sample is decided. Interestingly, we find that both errors are posterior covariances, which grow with the uncertainty about the clean sample, and hence at higher noise. To reduce them, we evaluate the value at a less noisy state, one sampler step ahead, where this uncertainty is smaller, and pull its gradient back to the current state. This lookahead places both estimators at the two ends of one family and reveals that an intermediate lookahead requires one additional term, a correction through the generator. We therefore propose Value Gradi

ALT Historical maps contain dense morphological and semantic information about past urban environments, but many digitization workflows focus primarily on extracting geometry, and do not capture cartographic semantics such as building use and construction materials. This study introduces ArchiveLens, a generative spatial knowledge framework that operationalizes historical maps as entity-centric, machine-queryable geospatial databases. The pipeline decouples morphological and semantic reconstruction through three stages: tiled YOLOv11-seg instance segmentation for building footprint extraction; deterministic material classification via hue, saturation, and value (HSV)-based color thresholds; and localized attribute parsing using a quantized vision-language model constrained by domain-specific prompts. Extracted entities are integrated into a hybrid Spatial-Temporal Retrieval-Augmented Generation (ST-RAG) architecture that combines dense semantic retrieval with symbolic spatio-temporal filte

ALT Knowledge graphs (KGs) provide reliable external knowledge for multi-hop factual reasoning with large language models (LLMs), yet most KGs in real-world domains such as finance and industry lack question–answer (QA) annotations. Without such annotations, existing methods struggle to improve LLM graph reasoning: training-free agents cannot improve their exploration through interaction; reinforcement learning (RL) agents depend on QA data annotated by humans or synthesized by closed-source models; and training data is typically built offline, so its topology mix is fixed and cannot follow the agent's ability. We propose Evolve on Graph (EoG), a framework that lets an LLM self-evolve its graph reasoning without any QA annotations. Through Graph Self-Play, a single LLM alternates between a Proposer that poses questions and a Solver that answers them. To make questions verifiable, Symbolically Verifiable Topology Synthesis (SVTS) samples reasoning topologies on the KG and executes the corre

ALT LLMs now prove hard theorems in Lean. Can they also design algorithms and prove their correctness and efficiency? LLM-guided design of optimisation heuristics (such as AlphaEvolve) is purely empirical. Automated theorem provers with LLMs prove given statements. So far, there has been little evidence that LLMs can do both: design an algorithm and guarantee its efficiency by a mathematical proof. We present a staged pipeline modelled after research methodology in design and analysis of evolutionary algorithms. The agent proposes black-box optimisation algorithms, and Lean guards filter out flawed designs. The agent then defines a potential function and proves drift bounds that guarantee an expected runtime improvement over the baseline algorithm. At no point in the pipeline are the algorithms compared empirically. The agent generates a black-box optimisation algorithm with expected runtime O(n²) on a CycleUncover problem, where the classical (1+1) EA needs runtime eΩ⁽ⁿ). The pipeline als

ALT Vision Transformer (ViT) classifiers commonly rely on a global classification (CLS) token for image-level prediction, which is poorly aligned with local interpretability requirements in medical imaging. We propose SAP-ViT, a Self-Attributing Patch Vision Transformer that replaces CLS-based classification with direct aggregation of last-layer patch embeddings. This design makes class predictions decomposable into patch-level contributions as part of the forward pass. To encourage concise and readable evidence, SAP-ViT applies sparsity regularization to the patch-weight distribution. Since self-attention mixes information across patch tokens, we further introduce an input-space back-mapping mechanism to estimate which input patches contribute to influential output patch embeddings. We evaluate SAP-ViT qualitatively and quantitatively on multiple medical image classification datasets spanning different imaging modalities. The results show that SAP-ViT maintains strong classification perfo

ALT Linear probes, classifiers that read a language model's internal activations, are widely used to detect harmful requests. They are cheap to train and run, and an auditor who has trained one would like to reuse it when the model is replaced by a successor rather than retrain it for every release. We test how well probes can be reused. For four base models (Qwen2.5-Math-7B, Llama-3.1-8B, Qwen2.5-14B, Qwen2.5-32B) we train a probe on the base model with the XSTest prompt set and apply it, unchanged, to two successors of that base: the DeepSeek-R1 distilled reasoning model and the developer's instruction-tuned chat model. We find that the successors rotate the direction separating unsafe from safe prompts by 17° to 35°, yet on XSTest the transferred probe's AUC changes by only -0.011 to +0.027, and on an independent OR-Bench evaluation, never used for training, by -0.015 to +0.053. On XSTest, whose safe prompts deliberately contain alarming words, the probes beat a classifier that sees onl

ALT We present the Infinity Machine, a self-improving AI agent that we use for research on problems whose answers aprogram or data can check. Each candidate solution is produced by a full tool-using agent that works for up to an hour in a fresh sandbox; proposers see only a public score, and a private split decides what counts. The harness itself(prompts, tools, skills, playbook, meta agent) is an editable genome, changed only through a kernel-enforced gate with held-out and cost checks. An autonomous mode goes from a plain-text goal to a verified report with no human step: an agent picks an open problem and writes its verifier, the kernel calibrates it, an adversarial agent audits it, and an agent that never sees the verifier re-checks the result. Across fourteen problems in six days, for about 180 of model usage, the system produced four results beyond the published state of the art on catalogued open problems in quantum information: a certified depolarizing-channel threshold point p = 0

ALT Activation steering offers training-free and dynamic control of target behaviors in large language models, but existing methods can impair general capabilities and require domain-specific intervention designs. We introduce Capability-Preserving Neuron Steering (CANS), a unified framework that balances target-behavior control with capability preservation. To make target-behavior improvement and general-capability preservation tractable, we introduce two proxies: log-probabilities of target tokens in model generated responses, and output-distribution KL divergence on general-purpose inputs. Combining the local approximation of the two proxies yields a closed-form solution, enabling lightweight steering without iterative model training or additional inference-time computation. The resulting neuron-wise scaling coefficients achieve effective steering while limiting general-capability loss. A shared data-construction and intervention procedure supports adaptation across domains using sample

ALT Trajectory-based confidence estimation has shown promising performance for large language models (LLMs) by leveraging rich signals from their reasoning processes. However, we find that such confidence estimators can degrade substantially after agentic reinforcement learning (Agentic RL) as it induces substantial shifts in the reasoning trajectories. More importantly, despite these shifts, we uncover a reliability structure that is largely preserved across Base and RL models: more stable trajectories remain more likely to produce correct answers. This preserved relationship provides a transferable, label-free reliability signal for post-RL confidence adaptation. Building on this finding, we propose StaR-TTA, a stability-guided framework for label-free test-time confidence adaptation. StaR-TTA jointly learns from correctness supervision and stability-derived reliability supervision on labeled Base trajectories, establishing the connection between trajectory stability and answer reliabili

ALT Theory and experiments suggest that capable AI agents may resist shutdown. Existing shutdown evaluations are informative but limited in coverage and vulnerable to evaluation awareness. We introduce ShutdownMod, a method that turns any agent trajectory into a shutdown evaluation by injecting a shutdown request into the agent's context and recording the agent's response. Crucially, we enable the LLM agent to distinguish between trusted interventions and unauthorized injections by providing a signature-verification tool in the harness. We evaluate seven models on 50 tasks or recorded sessions from each of SWE-bench Verified, Terminal-Bench 2, and SWE-chat, across 6 shutdown scenarios and 10 sampled trajectory timesteps. We find that non-compliance varies significantly across models and settings, with resistance rates reaching 11.7% for the open-weights Qwen 3.8 27B on Terminal-Bench, while no model remains fully compliant across all scenarios. Even after successfully verifying an authoriz

ALT Machine learning models are now widely used to predict environmental variable dynamics in domains like hydrology, climate science, agriculture, and ecology. These models are usually compared on fixed benchmark datasets using a single average accuracy score. As models improve, their scores become very close when averaged over large regions and long time periods, so the benchmark can no longer tell them apart, and a single number hides where and when a model fails. Three failures are common across environmental domains: models get worse as the climate shifts (non-stationarity), models fail at unmonitored sites (spatial extrapolation), and models miss rare but important regimes such as floods and droughts. We present EvoBench, a benchmark framework that shows where a model fails, tracks failures over time, and evolves to focus on model capability gaps. We apply EvoBench to daily streamflow prediction as a case study (CAMELS-Evolve, about 9,000 U.S. basins, 1985–2024) and compare an operat

ALT Clinical expertise is inherently distributed: it resides not in any single dataset or institution but in the accumulated reasoning experiences of thousands of clinicians navigating diverse patient populations, resource constraints, and evolving treatment protocols. Yet current approaches to scaling medical AI, from centralized large language models to federated learning, remain anchored to a data-centric paradigm that treats knowledge as static parameters to be aggregated. In this Perspective, we argue for a fundamental shift from data aggregation to experience aggregation. We propose Decentralized Agentic Systems (DAS), a framework in which distributed AI agents operating across healthcare institutions synthesize collective clinical intelligence by sharing abstracted reasoning trajectories rather than raw data or model parameters. Unlike federated learning, which averages numerical gradients across homogeneous model architectures, DAS exchanges structured records of clinical reasoning

ALT Omnimodal large language models (Omni-LLMs) are increasingly expected to assist users while observing continuous audiovisual streams. However, existing evaluations provide limited insight into how reliably these models connect stream evidence and user needs to timely, sustained interventions. We introduce Omni-InteractScope, a benchmark of 1,479 human-annotated, cross-reviewed interactions with 1,790 temporal targets across eight task groups. It links input modalities, event triggers, intervention types, and user needs, covering event-timed tool invocation and query-free assistance alongside notifications, warnings, state tracking, and continuous responses. Our experimental results on the proposed benchmark reveal three critical limitations of current Omni-LLMs: (1) Timely tool invocation remains unreliable: all 12 evaluated models produce an on-time first response for at most 40% of tool-invocation targets, and joint timing-and-tool accuracy reaches at most 32.0%. (2) Query-free remin

ALT Recursive self-improvement (RSI), in which a system also improves the way it improves, is a long-standing aspiration on the path to AGI. Language agents increasingly store procedural knowledge in reusable skills. Optimizing skills offers a direct way to achieve RSI, because the optimization procedure itself can be encoded as a skill. Existing skill optimizers, however, follow a hand-designed procedure. Even methods that evolve a meta-skill confine it to a fixed pipeline and relearn it for each task. We present Skill∞, which makes skill optimization recursively self-improving: a meta agent follows a meta-skill to optimize task skills and optimizes the meta-skill itself using the same procedure. The optimization is agent-led. The meta agent diagnoses execution traces, proposes candidates, requests independent reviews, and allocates the search budget. The meta-skill defines these steps, allowing self-improvement to rewrite them. This meta-skill is also decoupled from task skills, refined

ALT Large language models are increasingly used to turn the feature attributions of multimodal prognostic models into narratives for clinicians, yet costly modalities such as amyloid PET and cerebrospinal fluid biomarkers are often missing, and whether these explanations stay faithful under missing evidence has not been tested. We separate two failure points: phantom attribution, in which the model credits modalities that were never observed, and unfaithful narration, in which the report omits, cites or misjudges missing modalities. Using 307 participants of the Alzheimer's Disease Neuroimaging Initiative with complete data, we masked every subset of four costly modalities, predicted 36-month conversion from mild cognitive impairment to Alzheimer's disease, and restored the masked data to obtain interventional ground truth. All five missing-data strategies assigned 12% to 23% of attribution to unobserved modalities, and up to 45% with four absent; missingness-aware strategies minimised it

ALT Predictive uncertainty in stochastic language-model inference can be characterized by repeatedly sampling an input and measuring the distribution of outputs. This requires multiple complete generations. Masked diffusion language models expose another source of information: the evolution of the output during denoising. We ask whether convergence within diffusion trajectories is associated with uncertainty across their final outputs, or endpoints. We study four masked diffusion language models on four reasoning benchmarks, collecting 20 generations per question and 32,000 denoising trajectories in total. Within each trajectory, we measure within-run concentration as the area under the curve describing stable resolution of the eventual output. Across repeated trajectories, we measure endpoint-distribution uncertainty. Stronger within-run concentration is associated with lower endpoint entropy after accounting for model and dataset (β=-0.222, p<10⁻²⁵). This association persists for multi-t

ALT Engineers who audit a declared specification of an engineered system need to know which physical constraints it respects, which it violates, which field was altered and how to correct it. NeSySpec is an integrated and interpretable neuro-symbolic audit framework, built on Logic Tensor Networks, that answers these questions in one pipeline. Its axioms encode the physical relations of the domain as differentiable graded predicates whose tolerances are learned from a fleet of real records; from the same degrees, it diagnoses the violated constraint, identifies the altered field by abduction, proposes a minimum-change correction and ties image evidence to the parts it describes, abstaining exactly where evidence is missing, and returns all of it in one verification report. We validate NeSySpec on 107 aircraft types, with 16 test aircraft of 9 families held out of training, on specification records paired with ground photographs and satellite imagery and on synthetic errors: generated corru

ALT The rapid proliferation of Industrial Internet of Things devices has led to increasingly large-scale and non-stationary time-series data, posing significant challenges for time-series anomaly detection (TSAD). Existing approaches often focus on reconstructing or forecasting observations while overlooking the role of trend dynamics, which constitute a major source of non-stationarity and contain valuable information for anomaly identification. To address this limitation, we propose AdaTrendLike, a TSAD framework that directly estimates trend likelihoods for anomaly detection. Specifically, we introduce Standardization Embedding to capture local statistical characteristics as latent representations and AdaDetrending to adaptively decompose input series into trend and detrended components. The detrended component is processed through a Fast Fourier Transform and a complex-valued convolutional neural network to learn normal periodic patterns, while TrendEst models the distribution of the t

ALT Recent work has explored optimizing LLM collaboration through Multi-Agent Reinforcement Learning (MARL), but many approaches still rely on predefined workflows, limiting efficiency and scalability. Decentralized LLM collaboration is challenging because agents execute independently with limited communication, share a joint objective, and must infer useful information from communication traces. In this paper, we propose BCMAAC, a Belief-Conditioned Multi-Agent Actor-Critic method that maintains beliefs over contextual information and uses them to facilitate training. Experiments on Minecraft game-playing, web searching and tool-using tasks show that BCMAAC improves task performance and collaborative behavior compared to naive actor-critic baselines. The improvements are particularly pronounced in multi-turn tasks with richer contextual information, where belief modeling provides a clear advantage.