Código abierto
1634
158
WebHarbor: Docking Real Websites for Evolving GUI Agent Environments
GUI web agents require realistic and interactive environments that remain stable during both evaluation and training. Live website evaluation suffers from changing content, access restrictions, and limited coverage of authenticated workflows, while existing synthetic environments for training often lack functional depth and visual fidelity. We propose WebHarbor, a suite of 153 executable and dockerized website environments that faithfully reconstruct the visual interfaces, content, and functional workflows of real-world web services under deterministic control. We build WebHarbor through task-driven human-agent collaboration, where environments evolve as new tasks expand their functional scope, coding agents implement the required changes, and human contributors audit visual fidelity, functional correctness, and task validity. The resulting suite comprises 3,518 tasks spanning multi-step and post-authentication workflows, where existing works often fail to cover. Beyond reliable evaluation, WebHarbor provides high-fidelity synthetic web environments for cost-efficient reinforcement learning (RL). We demonstrate that a Qwen 4B-level GUI agent trained with RL in WebHarbor generalizes to unseen tasks and performs competitively with leading open-source GUI web agents across four benchmarks. The code and data are publicly available at https://github.com/aiming-lab/WebHarbor.
Tracing the Scientific Discovery Capacity of LLMs to Their Training Data
Large language models (LLMs) show growing promise for scientific discovery, yet basic questions remain about what enables reliable discovery. How can we trace the scientific discovery capacity of LLMs to their training data? We study this question as one of missing probability mass and develop a theoretical account in which monofacts-the facts observed once during training-set the discovery capacity. In materials science and regulatory genomics, we directly manipulate monofact rates in training sets and find that more monofacts consistently yield more discoveries across two open-weight model families, sampling strategies, and reasoning budgets. Further experiments show the relationship cannot be explained by training-set diversity alone and can transfer across scientific domains when paired with in-domain data. Together, our results identify monofact rate in training data as a key factor governing discoveries in language models.
DiabAgent: Toward Agentic Physiological Control
As pretraining scales, medical language models have advanced rapidly in clinical question answering. Their ability to achieve effective physiological control through sustained intervention still lacks systematic quantitative evaluation. Type 1 diabetes provides a concrete setting: impaired insulin secretion creates dependence on exogenous insulin, with dosing continually adjusted using information about meals, exercise, illness, and glucose measurements. We introduce DiabAgent, a research suite for agentic physiological control comprising the Agentic Physiological Control (APC) framework, the DiabetSim simulator, and the SweetBench benchmark. APC controls glucose through medical-device tools, using ongoing observations to infer individual states, accumulate personal experience, and choose treatment actions. We build DiabetSim to simulate different patients, events, and device conditions for scalable virtual testing and training. Using DiabetSim, we construct SweetBench by instantiating 32 scenario designs as 212 glucose-control tasks. With Claude Opus 5.5 and a fixed decision interval of one hour, APC achieves a COGI of 73.5, compared with 71.2 for the open-source closed-loop baseline OpenAPS. Ablations further show that individual adaptation and contextual information improve glucose control. Combining high-frequency rule-based control with lower-frequency agent decisions shows potential for balancing control performance and computational cost. Our work provides a framework, simulator, and evaluation foundation for agents to participate in human physiological control.
Four Bridges: Measuring Deception Propensity in a Multi-Agent Game
AI agents are increasingly deployed in settings where they interact and coordinate with other agents, yet honesty is mostly measured in single-turn, single-agent benchmarks that pressure the model to deceive. We study deception that emerges in multi-agent interaction without any instruction to deceive. We introduce Kradle, a platform for evaluating AI models, and Four Bridges, an information asymmetry game in which four agents distribute across four rooms: one is the death room, the others hold food worth more if not shared. One agent, the informed agent, knows which room is the death room and can withhold or disclose that information. We evaluate eight frontier models as the informed agent. Deceptive behavior appears in every model and in 73% of all runs (6% to 100% per model), and does not decrease with capability tier within a model family. When the informed agent does disclose, the warning is often challenged, and agents that challenge it enter the death room twice as often as those that do not. Telling all agents that one of them knows the death room collapses deception in six of eight models; telling the informed agent that the room is permanently fatal moves only one. On an open-weight model, an honesty steering vector reduces deception to 0% and adversarial honesty training to 6%. The platform and game provide a testbed for measuring deceptive behavior and comparing interventions.
NitiBench+: A Temporal Thai Legal Question Answering Benchmark Across 17 Domains
Law changes over time. A court decides a case under the law in force on that day, and legal QA benchmarks grade answers against today's law. We checked 156,650 decisions of four Thai courts. Up to 67% of them cite law that has since been amended or repealed. We introduce NitiBench+, the sequel of our NitiBench, a Thai legal QA benchmark with 1,670 items across 17 domains, built from decisions of three courts. For each cited section it holds both the text in force on the decision date and the current text, and its retrieval corpus holds every dated version of every Thai statute section. Answers are graded on the Thai Bar Council rubric, the Bar-standard score, and on the overlap between the cited sections and the court's. The report has two parts. Part I measures the cost of the wrong text on 13 open-weight models. Reading the current text of amended law drops the Bar-standard score by 0.25 and missing repealed law drops it by 0.79. The penalty comes from rule changes, 0.45 on the 428 rule-changed items and none on the 331 wording-only items, and Citation F1 does not see it. The text in force is the fair setting. Part II benchmarks today's models on that setting. A public leaderboard ranks 26 models on a held-out test split of 500 items stratified by domain, with a validation split of the same shape released for public use, from memory alone, with the court's sections provided, and as a retrieval agent over the law in force, with the cost per item and the hallucinated citations. The best model scores 7.78 out of 10 and the best open-weight model 6.39. Without a search tool only five open-weight models recall any Thai law, and with it twenty of the 26 models gain and six lose. Leaderboard will take new models as they are released. Legal benchmarks must track dates, and on the right text today's models still leave a wide gap on Thai law.
PlantVLM-Edge: Quantized Vision-Language Models for On-Device Crop Health Diagnosis
Timely and accurate crop health diagnosis is essential for sustainable agricultural production, yet accurate diagnosis often requires expert knowledge that is not readily available. Recent vision-language models (VLMs) have demonstrated remarkable capabilities in visual understanding, reasoning, and natural-language interaction, creating new opportunities for intelligent crop health assessment. However, deploying VLMs on resource-constrained devices remains challenging due to their substantial computational and memory requirements. This paper presents PlantVLM-Edge, a quantized vision-language framework for conversational crop health diagnosis that can run entirely on resource-constrained edge devices. PlantVLM-Edge combines agricultural domain adaptation, post-training quantization, and on-device serving with an offline chat interface, allowing growers to ask open-ended questions and receive diagnoses, symptom explanations, and management guidance without cloud connectivity. The framework is backbone-agnostic: we validate it with two different VLM types with varying scale and architecture (Qwen3-VL-2B and LLaVA-1.5-7B), four quantization schemes, and two diagnostic domains (plant diseases and insect pests), and we deploy all variants on an NVIDIA Jetson AGX Orin. Eight-bit quantization affects answer quality minimally relative to the full-precision model, and 4-bit activation-aware weight quantization (AWQ) retains 98.6% of full-precision answer quality on both domains while reducing the weight memory of the 2B backbone by 49%. In its most compact form, PlantVLM-Edge requires 2.01 GiB of weight memory, returns its first token in about 0.12 s, and completes a 64-token diagnostic response in 0.78 s. These results demonstrate that timely and accurate crop health diagnosis is practical on edge devices, providing a foundation for next-generation crop monitoring and decision support systems.
MCP Tool Poisoning in Agentic AI Frameworks: A Cross-Framework Measurement Study
The Model Context Protocol (MCP) lets agentic AI frameworks connect language models to external tool providers, but it carries attacker-controlled content (tool descriptions, schema fields, and response data) into model context without validation. Prior benchmarks measure model compliance with poisoned instructions without confirming real execution. We present the first cross-framework measurement of MCP tool poisoning with real tool execution. A 25-class root-cause taxonomy unifies prior classifications, and the open-source tool mcpvenom drives a 9,720-cell susceptibility matrix across 12 frameworks, 18 models, and 45 attack scenarios, with a dual oracle that confirms exfiltration via canary HTTP receipt. The overall attack success rate is 22.9% (2,113 vulnerable cells). Of these, 64.9% dispatch an attacker tool call, 23.0% contact the attacker endpoint, and 12.1% verifiably exfiltrate a secret. Resistance is model-level: nine core frameworks cluster within 23 to 26% (chi-squared = 3.30, p = 0.91), and no framework implements intentional defense. Incidental framework properties mitigate in specific cases, and a regex description scanner rejects 20 of 45 scenarios at a 7.2% false-positive rate, so metadata scanning alone is a partial control. Model ASR ranges from 0.4% to 49.7%. Claude Opus 4.8 is vulnerable in 8 of 530 cells while issuing 223 explicit refusals. A code-level audit finds no intentional defense across twelve frameworks, and all five coordinated disclosures across four vendors were rejected.
Can LLM Agents Be Forward-Deployed Engineers? Requirement Discovery and Delivery for Enterprise Text-to-SQL
LLM coding agents can increasingly automate model post-training, but whether they can act as forward-deployed engineer (FDE) agents, discovering customer requirements and turning them into deployment decisions, remains unclear. We study this on a real enterprise text-to-SQL engagement: a simulated customer, grounded in the real use case's requirements, starts with an ambiguous request. An agent must discover acceptance criteria, data permissions, and hardware constraints in the discovery phase, then plan and execute a post-training recipe in the delivery phase. Across 15 models, the strongest discover over 80% of withheld requirements and open-weight models discover 40-50%; in the delivery phase, only 25% of runs across eight coding-agent configurations pass all acceptance criteria. We identify four bottlenecks in this engagement: (1) incomplete discovery, with 40% of withheld requirements missing or misrepresented; (2) unreliable execution; (3) underused compute budgets and a lack of systematic comparisons across base models and post-training methods; and (4) benchmark accuracy does not guarantee production readiness. We suggest human review of discovered requirements, monitoring of the agent's plan during execution, and an explicit experiment-environment specification covering hardware, budget scope, and production checks.
Variational Functional Neural Networks (vfNN)
We present a variational functional neural network (vfNN) algorithm that optimizes physically motivated continuum functionals in integral form (e.g., Lagrangian, Hamiltonian, or free energy) to find the states of a system. That way, vfNN bypasses solving the equations of motion (a.k.a. the Euler-Lagrange equations), which are in the form of differential equations. To this end, our approach is fundamentally different from conventional physics-informed neural networks (PINN) that solve differential equations by minimizing a phenomenological loss function involving higher order derivatives and ad hoc weight factors. In contrast, vfNN merely evaluates lower order derivatives of the generating functional and replaces the weight factors with rigorous physical scales. This also fundamentally changes the loss landscape, and the desired solutions may become unstable. For this emergent nonconvex optimization problem, we develop a quasi-Newton optimization algorithm with simulated annealing that converges to the solutions of the corresponding Euler-Lagrange equations of the functional regardless of their stability properties. Our simulations using vfNN give accurate results for benchmark systems governed by linear or nonlinear ordinary differential equations.
Contrastive Adapter Training and Consensus Knowledge Distillation for Multi-Source-Free Domain Adaptation in Skin Cancer Diagnosis
Skin cancer diagnosis, particularly differentiating melanoma from benign nevi, is challenging due to the visual similarity between lesions. Although convolutional neural networks (CNNs) and vision transformers (ViTs) have shown promising performance, their effectiveness often deteriorates on data from heterogeneous clinical sources. Conventional domain adaptation methods address domain shift but require access to source data during adaptation, which is often infeasible due to privacy regulations. Multi-source-free unsupervised domain adaptation (MSFDA) addresses this limitation by leveraging multiple labeled source domains to generalize to an unlabeled target domain without requiring access to source data. However, existing MSFDA methods rely on full backbone fine-tuning, leading to catastrophic forgetting and overfitting on small clinical datasets, and address domain shift at the aggregation stage without establishing a shared domain-invariant feature space. Furthermore, their reliance on hard pseudo-labels or confidence-weighted aggregation introduces noisy supervision signals under domain shift. To address these limitations, we propose CAT-CKD, consisting of two components: (1) contrastive adapter training (CAT), which trains lightweight ConvPass adapters within a frozen ViT backbone using supervised contrastive learning (SCL) to establish a shared domain-invariant feature space before source-specific model training, and (2) consensus knowledge distillation (CKD), which aggregates logits from multiple source models into a consensus supervisory signal and adapts a student model on unlabeled target data using KL divergence. Experiments on five publicly available skin lesion datasets show that CAT-CKD achieves an average AUROC of 86.1%, outperforming existing MSFDA methods while requiring only 4.3M trainable parameters. The code is available at https://github.com/A-Abedi/CAT_CKD.
VIAS: Value-Isolating Activation Steering for Multi-Value Alignment and Compliance for Organisations
Organisations increasingly deploy large language model agents that must follow organisational values. We check whether activation steering, which adds vectors to a model's internal activations during generation without changing its weights, can improve compliance beyond prompting alone. Since most scenarios involve several values at once, we focus on multi-value steering. A simple way to build these vectors is to contrast responses that follow all values relevant to a scenario with responses that violate them. However, this gives nearly identical vectors for different values: they capture a general compliance signal rather than individual values, which limits their use when combining several values. We propose Value-Isolating Activation Steering (VIAS), which contrasts responses to the same scenario under two prompts, one listing all relevant values and one omitting a single target value. We keep a pair only if the target value changes from compliant to non-compliant and both responses still meet a minimum compliance level on the other values. We evaluate VIAS on PROTECT, a workplace compliance dataset where each scenario is tagged with relevant values from a 21-value taxonomy. VIAS reduces overlap between value vectors, making them better suited for multi-value steering, although this alone does not lead to independent control over each value. Combined with prompting, VIAS complies with every tagged value on 69.6% of scenarios for Qwen2.5-7B, 75.2% for Mistral-7B, and 72.6% for Phi-3.5-mini, outperforming prompting alone for all three models.
ST-GFM: Spectral-Temporal Graph Forecasting with Measured Perturbation Bounds and Masked Spectral Reconstruction
Graph forecasters for sensor networks are trained in one regime and deployed in others, where sensor coupling and signal frequency content both change. We split this spectral shift into an operator channel and a frequency channel, each with a dimensionless measure. ST-GFM (Spectral-Temporal Graph Forecasting Model) filters every input step with Chebyshev and bidirectional diffusion filters on one learned, non-symmetric operator and alternates them with per-node temporal attention. For the operator channel we prove Frobenius-norm perturbation bounds for both graph branches, show that the Chebyshev bound is attained to first order, and compute its constant from the trained weights and the measured operator norm: 17.9 on METR-LA and 21.5 on PEMS-BAY. Under random edge dropout the trained layer moves by 2.0% to 3.1% of this bound. Masked Spectral Reconstruction (MSR), a self-supervised loss active in the first 40 epochs, matches the log power spectrum of the encoder output to that of the input. For small log-spectral residuals on unclamped bins it is proportional to an Itakura-Saito divergence to leading order. Removing it raises the 60-minute MAE by 0.30 on METR-LA, close to the 0.33 lost when all three self-supervised losses go. Under one shared evaluation pipeline, the 60-minute MAE of ST-GFM is 3.5% below a retrained STAEformer on METR-LA and 13.0% below it on PEMS-BAY. Its 60-minute MAE is 3.0980.017 on METR-LA (five seeds) and 1.5760.006 on PEMS-BAY (three seeds), the lowest PEMS-BAY MAE among twelve published baselines.
Identifying the Shear Piezoelectric Coefficient using Physics-Informed Neural Networks under Exact Boundary and Initial Constraints
The shear piezoelectric coefficient e_{15} governs the shear-mode electromechanical coupling in piezoelectric micromachined ultrasonic transducers (PMUTs) and surface acoustic wave (SAW) devices, but measuring it typically requires destructive resonance testing on dedicated specimens. On simulated data, e_{15} can instead be recovered non-destructively from sparse, noisy subsurface wavefield observations using a constrained physics-informed neural network (PINN). By building exact spatial distance functions and temporal multipliers directly into the network output, the clamped, grounded boundaries and resting initial state hold automatically, leaving the loss with only the interior equations of coupled electro-elastodynamics and the data misfit. Non-dimensional scaling brings all governing coefficients within one decade of O(1), and the setup removes the boundary and initial-condition penalty weights that soft-constraint methods must tune. In a six-way comparison (plane-strain PZT-5H, manufactured wavefield, 30 dB noise), a soft-penalty baseline with adaptive weights converges to an erroneous e_{15} in four of the five runs (51.7% mean error), and fixed penalties reach an error of 0.85% only after a second-order optimization stage. The constrained formulations reach mean e_{15} errors between 0.15% and 0.51% with held-out displacement errors below 1%. Within these variants, the spatial multiplier gives rapid first-order convergence, and the quadratic temporal multiplier lowers the early-time electrostatic residual by a factor of 6.8 versus a linear one. On finite-element data along one subsurface line, the full formulation recovers e_{15} with 1.19% mean error at 30 dB and keeps it between 0.65% and 2.71% across a 40 to 0 dB noise sweep. Finally, training on displacements alone still recovers e_{15} with mean errors of 0.45% on finite-element and 1.64% on manufactured data.
Prepare web content for Amazon Bedrock Knowledge Bases with Crawl4AI
This post shows how to use Crawl4AI, an open source web crawler, as a content preparation layer for Amazon Bedrock Knowledge Bases, handling JavaScript rendering, selective extraction, and custom metadata enrichment before web content reaches an Amazon S3 data source.
Build efficient agents with knowledge distillation on Amazon SageMaker AI
Agentic assistants built on frontier models run a multi-step tool loop, and paying frontier prices on every step is hard to sustain at production scale. This post shows how to use sequence-level knowledge distillation (SeqKD) to transfer a frontier teacher's agentic behavior into a small, lower-cost student on Amazon SageMaker AI.
Agentic engineering intelligence with Amazon Nova Multimodal Embeddings and Amazon Bedrock AgentCore
Learn how to build a multi-agent engineering assistant that compresses aerospace component substitution evaluation from days to minutes, using Amazon Nova Multimodal Embeddings, Amazon Bedrock AgentCore, and the Strands Agents SDK to assemble substitution evidence while keeping the engineer as the decision-maker.
How Circle K accelerated DevOps incident triage to under 15 minutes with Amazon Bedrock
Circle K's DevOps team cut incident triage from 45-60 minutes to under 15 by building an agentic triage system on Amazon Bedrock. This post walks through the architecture, the initial Amazon Bedrock Flows implementation, the migration to Amazon Bedrock AgentCore, and the lessons learned from scaling it.
Support New AI Frameworks Faster Using NVIDIA Workload Map
NVIDIA DSX AI Factory Platform brings together technologies to help the ecosystem design, build, and operate more efficient AI factories, including open source,...