Artificial Intelligence
A Governed Agent Architecture for Natural-Language Dashboard Generation over Enterprise DataBrain
Yiou Wu, Zezhi Tang, Ningwei Bai et al. (+1 more)
We present a DataBrain-specific prototype that turns natural-
language analytic requests into inspectable dashboards. An installable Codex plugin and standalone
Agent Runtime combine schema-constrained model decisions with typed tools, persistent state, and deterministic Hooks for approval, audit, checkpointing, recovery, and failure handling. The pipeline resolves entities, discovers metadata, enforces read-only SQL, composes dashboards, and applies static checks, dynamic preflight, and browser inspection. The model proposes actions while deterministic software controls execution and records state transitions. We evaluate the workflow on frozen real-DataBrain tasks and controlled Hook faults. Strict success was 6/8 on metadata and SQL tasks: metadata selection passed 4/4, all four SQL tasks met semantic criteria, and 2/4 met the exact output-column contract. The final release passed 4/4 single-panel dashboard tasks, one two-panel task, and one existing-dashboard refinement; a parameterised task exceeded its step limit. All ten fault scenarios met their specified outcomes without unapproved external side effects. Model inference accounted for over 97% of observed runtime in every reported group. These small, DataBrain-specific results do not establish production readiness, general text-to-SQL accuracy, or an efficiency advantage over manual dashboard construction.
Read the full abstract ▾
All authors Yiou Wu, Zezhi Tang, Ningwei Bai, Liuhaichen Yang
Artificial Intelligence
GAMEGO: Training Game-Dev Agents with Synthetic Trajectories Anchored in Real-World Assets
Haoyue Yang, Jingyao Li, Zhengfan Wu et al. (+3 more)
Recent advances in Large
Language Models (LLMs) have demonstrated remarkable capabilities in web front-end execution, with browser-based game generation emerging as a particularly prominent frontier. While previous efforts frequently rely on complex multi-turn workflows or focus on static game evaluation
benchmarks, this work targets direct end-to-end real-world game synthesis driven by coding
agents. However, generating complex games directly from sparse user queries often forces coding
agents to make underspecified assumptions, yielding incomplete mechanics, disconnected gameplay flows, and limited visual aesthetics. To resolve this issue, this paper presents GameGo, a scalable framework that systematically transforms brief game seeds into comprehensive Product Requirements Documents grounded in industry game-development practices. To retain core gameplay constraints without restricting design exploration, GameGo uses task-specific dynamic compression to maximize information density while preserving instruction following. Based on this pipeline, GameGoData is constructed with 55,060 development trajectories across 2D, 2.5D, and 3D games, alongside GameGoBench, a
benchmark comprising 124 diverse game queries. Training GameGoCoder on GameGoData yields a model that outperforms matched baselines and is comparable to frontier models across gamedev
benchmarks. All code,
datasets, and models will be made publicly available.
Read the full abstract ▾
All authors Haoyue Yang, Jingyao Li, Zhengfan Wu, Jing Liu, Xuanle Zhao, Kang Liu
Artificial Intelligence · Cross-listed from cs.CL
Trajectory-Derived Confidence for Reliable, Resource-Aware Clinical Text-to-SQL Agents
Mincheol Daniel Song, Joshua Ward, Jake Jung et al. (+1 more)
LLM
agents for clinical text-to-SQL applications reason autonomously over multiple steps but cannot assess whether their own
reasoning or outputs can be trusted. In high leverage applications such as healthcare, this presents a critical risk where system mistakes can be costly. These reliability failures are also resource failures: an incorrect
reasoning trajectory spends computation budget on outputs that must be discarded. We introduce Sentinel, a trajectory-derived, classifier-based confidence layer that analyzes an
agent's
reasoning, code and database outputs to decide at three points whether to stop: refusing unanswerable questions before the
agent runs, halting doomed trajectories mid-run, and withholding untrustworthy answers at delivery. Here, utilizing Chow's rule, we optimize decisions under the EHRSQL shared task's Reliability Score, which penalizes incorrect answers given a utility weighting, and find on the
benchmark EHRSQL that Sentinel raises this score from +0.08 to +0.24 when mistakes have a low utility weighting, well above the +0.03 earned by refusing every question, with delivered-answer accuracy rising from 54% to 69% as coverage falls from 84% to 55%. At higher stakes the
agents we test rarely answer reliably enough to deliver, and Sentinel detects this on its own, abstaining to that same +0.03 where the unmonitored
agent scores -3.02. The same stopping decisions cut computation: at low stakes, where the system still answers, Sentinel eliminates 13-28% of
agent steps at little or no reliability cost.
Read the full abstract ▾
All authors Mincheol Daniel Song, Joshua Ward, Jake Jung, Guang Cheng
Artificial Intelligence · Cross-listed from cs.LG
Dual-Scale Relational Graph Transformers for Ecosystem-Aware Fraud Detection
Mohsen Nayebi Kerdabadi, Xinrou Li, Yao Xiao et al. (+2 more)
Account takeover (ATO) fraud is a growing threat to digital banking, requiring effective detection while minimizing friction for legitimate customers. Production systems predominantly rely on tabular models that score sessions in isolation, discarding the relational structure of the underlying interaction network. Although graph-based models exploit relationships among sessions and network entities, they primarily reason over local neighborhoods and therefore capture only part of the problem: fraud risk depends jointly on the local relational structure surrounding a session and the evolving global state of the fraud ecosystem. We present HERMES (HEterogeneous Relational Micro–macro graph
transformer Encoder for high-risk Sessions), a dual-scale architecture that jointly models these complementary scales of information. Micro-GT captures local heterogeneous graph structure through structured, relation-aware
attention over a temporally safe session neighborhood. Complementing this local representation, Macro-GT models ecosystem-level context using non-anticipative
climate tokens that summarize fraud dynamics, platform shifts, and infrastructure reuse, together with adaptive class prototypes that track representative fraud and benign session patterns over time. Evaluated on more than 130 million high-risk transaction sessions from a leading U.S. financial institution, HERMES consistently outperforms production and strong graph-based baselines, achieving a 44.44% relative reduction in customer friction and a 24.66% relative improvement in fraud recall over the production system. Ablation and temporal-stability analyses further demonstrate complementary gains from local relational modeling and global ecosystem context across changing fraud regimes.
Read the full abstract ▾
All authors Mohsen Nayebi Kerdabadi, Xinrou Li, Yao Xiao, Zijun Yao, Xin Sun
Artificial Intelligence · Cross-listed from cond-mat.mtrl-sci
Agentic Resource Allocation for Batch Multi-Objective Bayesian Optimization in Autonomous Materials Discovery
Robert Robinson, Shakti Prasad Padhy, Sushant Sinha et al. (+6 more)
The discovery and development of advanced materials is a challenging process constrained by the high time and monetary costs of synthesis, processing, and characterization. The underlying design spaces can be enormous, often with multiple competing objectives.
Bayesian optimization (BO) provides a principled approach for efficiently navigating such spaces, but most workflows rely on fixed exploration-exploitation policies that lack the capacity to adapt to shifting constraints in dynamic campaigns typical of self-driving laboratories. In this work, we develop a multi-objective BO framework for
alloy design under resource constraints, benchmarking strategies for adaptive policy tuning at each iteration. Our evaluation covers a septenary refractory high-
entropy alloy (RHEA) system focused on maximizing melting temperature and minimizing density, and an Fe-Co-Ni-based soft
magnetic alloy system targeting saturation magnetization, coercivity, and hardness. We compare an exploitation-focused strategy, a fixed mixed exploratory/exploitative policy, and two distinct LLM-based adaptive strategies with different approaches to batch allocation and campaign signal interpretation, evaluated across baseline and mid-campaign resource event conditions including budget reductions, timeline cuts, and combined disruptions. Our results show that mixed allocation strategies accumulate substantially more mutual information than the exploitation-focused baseline at a proportionally smaller cost to hypervolume and optimization speed, with adaptive strategies outperforming a fixed-mixed allocation policy by adjusting their allocation in response to both evolving campaign statistics and resource constraints. These findings suggest that adaptive resource allocation offers a favorable tradeoff for materials discovery campaigns in reducing predictive uncertainty on Pareto-optimal compositions.
Read the full abstract ▾
All authors Robert Robinson, Shakti Prasad Padhy, Sushant Sinha, Sk Md Ahnaf Akif Alvi, Juan Florez Coronel, Brent Vela, Trevor Hastings, Douglas Allaire, Raymundo Arróyave
Artificial Intelligence
Auditing Pairwise Equivalence Judgments: Self-Critique Effects and Diversity Measurement in Multi-Agent Hypothesis Generation
Ji Young Byun, Anthony Hu, Jesse Rogers et al. (+3 more)
Multi-
agent systems built on large
language models (LLMs) are increasingly applied to scientific discovery and hypothesis generation. Both the effect of refinement and the diversity of the delivered set are hard to interpret before experimental ground truth exists, and both are typically reported by deciding whether pairs of generated hypotheses describe the same underlying mechanism. We study two evaluation questions that rest on this pairwise equivalence judgment: (1) how much self-critique changes delivered hypotheses beyond run-to-run variability, and (2) how the equivalence rule used to group hypotheses affects measured diversity. Across four proprietary instances, we hold opening hypotheses fixed, rerun the downstream workflow with 0, 1, and 5 critique rounds, and score matched hypothesis pairs with an LLM-as-a-judge. Relative to matched same-depth reruns, moving from 0 to 1 round produces 34.5 percentage points (pp) of additional mechanism-level divergence, whereas 1 to 5 rounds adds 1.3 pp. We then compare three equivalence rules: term frequency–inverse document frequency (TF–IDF) similarity, dense
embeddings, and the same LLM-as-a-judge. We construct controlled hypothesis pairs that either preserve the causal explanation through wording or biological-terminology changes, or replace one component of the causal chain while holding the rest fixed. All three rules are invariant to meaning-preserving edits, but when the initiating event is replaced, the LLM-as-a-judge identifies 83% of valid pairs as different mechanisms, versus 0% for TF–IDF and 8% for
embeddings; varying only the rubric that defines same mechanism moves this figure from 38% to 96%. Together, these results show that pairwise equivalence judgments are a measurement choice: how mechanism equivalence is defined affects both the estimated effect of self-critique and the measured diversity of generated hypotheses.
Read the full abstract ▾
All authors Ji Young Byun, Anthony Hu, Jesse Rogers, Roujia Wang, Manasa Kesapragada, Falgun Shah
Artificial Intelligence · Cross-listed from cs.CL
PB-GRPO: Learning Socially Adaptive LLM Agents from Persona-Driven Simulation with Preference-Batched GRPO
Jingquan Wang, Jun Yin, Xu Han et al. (+3 more)
Building LLMs that behave well socially, not merely correctly, requires Building LLMs that behave well socially, not merely correctly, requires more than producing locally helpful responses. A socially competent
agent must infer users' unstated goals, respect their preferences, and adapt as the conversation unfolds. These behaviors are inherently multi-turn and social, making them hard to optimize: real interaction data is scarce, and user preferences are typically latent rather than directly observable. To address these challenges, we build on a persona-driven social simulation environment (consisting of a persona library, LLM-based user simulators, and a user-satisfaction scoring system ranging from [0, 1]), to introduce preference-batched GRPO (PB-GRPO), a post-training
algorithm that learns socially adaptive policies from conversation-level feedback. Compared to vanilla GRPO, PB-GRPO computes advantages using a normalization estimated across a bucket of users with similar preferences, stabilizing training across a diverse social population. Empirical evidence shows that PB-GRPO improves models' social behavior over strong reinforcement learning baselines in our simulated environment.
Read the full abstract ▾
All authors Jingquan Wang, Jun Yin, Xu Han, Yongsheng Mei, Jie Hao, Bin Guo
Artificial Intelligence
InvestigationWorlds: An Agentic Environment for Legal Investigation
Albert Yu Sun, Andrew Benard, Sil Hamilton et al. (+8 more)
We introduce InvestigationWorlds, an agentic environment for legal investigation. We build on an underused artifact of U.S. civil litigation: the summary judgment motion. This motion relies upon a record composed of real evidence exhibits, and results in a court-adopted hypothesis that is treated as ground truth for the purposes of deciding the motion. Each environment is built from a real U.S. Federal Court case retrieved from Public Access to Court Electronic Records (PACER) and augmented by an attorney-validated generation pipeline that synthesizes role-tagged documents around the original record. The resulting corpus admits multiple coherent factual readings, only one of which matches the court-adopted hypothesis. Evaluating on 100 cases, we find
agents often commit to incorrect hypotheses despite retrieving relevant evidence, struggling to distinguish the court-adopted hypothesis from alternative hypotheses.
Read the full abstract ▾
All authors Albert Yu Sun, Andrew Benard, Sil Hamilton, Anna Teresita A. Marcelo, Yong Jae Kim, Carl-Leander Henneking, Rundong Hu, Yuhong Wang, David Mimno, Bishan Yang, Igor Labutov
Artificial Intelligence · Cross-listed from cs.CL
Representational Control over Self-Report & Behavior Coherence in LLM Risk-Taking
Rafal Kocielnik, Peiyang Song, Pengrui Han et al. (+4 more)
Self-report is an appealing low-cost probe of an LLM's dispositions, but recent work finds only selective agreement between what models report and how they behave. Prior accounts establish these patterns by prompting black-box LLMs, leaving open whether the gap is a prompting artefact or a fact about how the underlying constructs are represented internally. We investigate risk-taking, a consequential dimension of agentic decision-making, using activation steering to measure self-report and behavior under the same internal intervention. We survey nine steering-vector extraction methods spanning task-specific directives, the model's own task behavior, and dispositional descriptions at two granularities, evaluated on two behavioral tasks and two psychometric instruments across four open-weight LLMs. We find that (1) a shared internal intervention does not ensure shared responsiveness: directions built from trait descriptions move self-report but leave behavior at chance, directions built from the model's own task choices do the reverse, and only task-specific directives reach both, weakly. (2) Diagnosis dissolves that exception: removing surface confounders leaves the directives only 32% of their behavioral effect. The two channels are otherwise steered by near-orthogonal directions, each channel reachable by several independent constructions. (3) An intervention composing one behavior-moving and one self-report-moving direction moves both together; flipping one sign sets them in opposition, with reported and enacted risk pointing in opposite directions on 69-89% of flipped compositions in all models. These findings move the self-report-behavior relationship from a black-box observation to a representational one that can be inspected and controlled, motivating representational checks alongside behavioral evaluation.
Read the full abstract ▾
All authors Rafal Kocielnik, Peiyang Song, Pengrui Han, Myrl G. Marmarelis, Ramit Debnath, Dean Mobbs, R. Michael Alvarez
Artificial Intelligence · Cross-listed from cs.CY
Agent Reliability Profiles in Financial Services
Mike Hsu, Medha Bankhwal, Béatrice Moissinac et al. (+3 more)
AI
agents can take actions. At times, those actions can go beyond what is intended.
Agent reliability can be defined as assurance that an
agent will stay within intended bounds and operate within limits. Today, there is no shared framework or
language for describing, validating, and benchmarking the reliability of agentic deployments in financial services. This makes it difficult for financial institutions, vendors, and regulators to assess and trust
agents at scale, thus limiting the pace of development and adoption. A standardized, shared representation of
agent reliability would fill the gap. This paper introduces the
Agent Reliability Profile, a per-
agent unit of assurance evidence for
agent deployments in financial services. Each Profile records a bounded, falsifiable claim, this agentic system reliably functions within its operating boundary. We define "operating boundary" as an
agent having; (1) a defined autonomy tier, (2) a defined operational design domain, (3) defined classes of action, and (4) a defined control envelope. Production assurance progresses through three levels while the Profile schema remains constant: a Profile Builder compiles a Level 1 Asserted Profile from institutional evidence, a Profile Validator tests the deployment in its own environment to produce a Level 2 Validated Profile, and operation of the same tests by a qualified independent assessor produces a Level 3 Verified Profile. Separately a Benchmarked Profile reports results comparable across institutions under reference conditions. We describe the architecture, the artifact, the assurance ladder, the comparability flag, associated tools, an evaluation methodology, applications for financial institutions and supervisors, limitations, and a staged implementation program.
Read the full abstract ▾
All authors Mike Hsu, Medha Bankhwal, Béatrice Moissinac, Kevin Werbach, Lukasz Szpruch, Bennett Hillenbrand
Artificial Intelligence
CUAWright: A Minimal Unified Interface for Digital Agents
Yadong Lu, Theodore Lee, Yifei Li et al. (+5 more)
The prevailing approach to computer-use
agents couples a model with a domain-specific harness: a browser or desktop environment equipped with human engineered tools that are fixed before task execution. As models' coding capabilities improve, the GUI native and static harness prevents them from direct programmatic operation on system state, as well as flexible construction of tools. To this end, we introduce CUAWright, a minimal terminal harness of roughly 3K lines of code that uses bash commands as its sole action interface, and a file system as its evolvable space for dynamically creating tools and managing the context. We conduct comprehensive experiments across a wide range of digital tasks, and demonstrate that by giving the
agent a minimal, programmable interface, it achieves substantially stronger results compared to their GUI or hybrid CLI interface across a wide range of tasks. On OSWorld 2.0, CUAWright delivers a 33.2% relative improvement in partial reward while reducing estimated cost by 37.5% compared with the published GPT-5.5 baseline. On Online-Mind2Web and the long horizon Odysseys
benchmark, CUAWright substantially outperforms GUI native harness by 4.7% and 44.0% in success rate, respectively. Furthermore, we found the gains extend to CAD applications that require accurate visual understanding and CLI interaction: on CADGenBench and BenchCAD, our unified harness yields 8.1%-41.6% relative improvements over other CLI based harnesses with GPT-5.5. Together, these results suggest digital environments are far more programmable than their GUI interfaces imply, and a minimal terminal-focused harness is the key for better performance and efficiency.
Read the full abstract ▾
All authors Yadong Lu, Theodore Lee, Yifei Li, Lawrence Keunho Jang, Tianci Xue, Yu Su, Huan Sun, Ahmed Hassan Awadallah
Artificial Intelligence
The Independence Prior of SAEs Fragments Visual Concepts
Tommaso Mencattini, Giorgos Nikolaou, Donato Crisostomi et al. (+4 more)
Sparse Autoencoders (SAEs) decompose model activations into sparse combinations of interpretable dictionary atoms. Although SAEs are grounded in the Linear Representation Hypothesis (LRH), their objective smuggles in an additional prior: concepts across patches are treated as independent, an assumption clearly violated by natural images and by the activations they induce. We therefore specialize LRH to
vision through the Markov-Field Linear Representation Hypothesis (MFLRH), which adds the missing spatial dependencies to the LRH assumptions. We thus propose Spatial-SAE as an amortized MAP
estimator under the MFLRH. Spatial-SAE consistently outperforms standard SAEs in concept recovery and interpretability. Across four variants, it achieves a 96% average win rate on synthetic concept recovery and improves interpretability on DINOv2 activations, at a reconstruction cost concentrated in high spatial frequencies.
Read the full abstract ▾
All authors Tommaso Mencattini, Giorgos Nikolaou, Donato Crisostomi, Thomas Fel, Francesco Montagna, Emanuele Rodolà, Francesco Locatello
Artificial Intelligence · Cross-listed from cs.LG
SEER: Self-Evolving Event Reasoning and Retrieval for Time Series Forecasting
Mingtian Tan, Palash Goyal, Mihir Parmar et al. (+6 more)
Real-world time series are frequently driven by exogenous events and structural shifts, rendering conventional forecasting based solely on historical numerical observations insufficient. While
language models can retrieve external news, standard
retrieval-augmented approaches struggle with high noise, missing signals, and an inability to reason causally about event impacts. We propose SEER (Self-Evolving Event
Reasoning and
Retrieval), a closed-loop framework that dynamically optimizes event conditioning for time series forecasting. SEER translates prediction errors into two decoupled feedback mechanisms: (i) a reflective
retrieval memory that refines subsequent search queries and filters spurious noise, and (ii) a persistent causal knowledge base that distills transferable domain dynamics. SEER enforces strict chronological boundaries across both event
retrieval and reflection, preventing look-ahead bias and data leakage. Across six volatile time-series
benchmarks, SEER consistently outperforms state-of-the-art time series foundation models and
language model baselines.
Read the full abstract ▾
All authors Mingtian Tan, Palash Goyal, Mihir Parmar, Sarkar Snigdha Sarathi Das, Chun-Liang Li, Nanyun Peng, Thomas Hartvigsen, Jinsung Yoon, Tomas Pfister
Artificial Intelligence · Cross-listed from cs.LG
Robust blind unmixing: A geometric approach to overcoming basis variation
Dumitru Mirauta, Vladimir V. Gusev, Michael W. Gaultois et al. (+2 more)
Signal separation problems are common in science. A prominent example of this occurs during the use of diffraction or spectroscopy to identify the individual components of a mixture by measuring it. In the simplest case, the measured signal is a linear combination of basis patterns corresponding to the constituent parts. The unmixing problem is to infer all or some of these basis patterns and abundances of components from measurements of distinct mixtures. One of the core challenges of this task is the variation of the basis from mixture to mixture due to noise and the exact physics of the measurement process. This is usually addressed with tailored model-based and parametric methods that are then limited in use to specific application domains by the nature of the assumptions made. We propose a novel
geometric approach to unmixing problems which views the generation of data during measurement through a metric space lens, thereby shifting the focus from parametrised models to a general relationship between basis transformations and the corresponding geometry. We take advantage of the optimal transport distances to capture commonly occurring basis variations, and use minimisation of in-class variance of candidate solutions to drive the optimisation. We pay special
attention to the one-dimensional case due to its practical importance and availability of efficient distance and transport map routines. The effectiveness of our approach is demonstrated on a range of unmixing tasks using random Gaussian mixture models, simulated powder X-ray diffraction, and laboratory hyperspectral imaging
datasets.
Read the full abstract ▾
All authors Dumitru Mirauta, Vladimir V. Gusev, Michael W. Gaultois, Matthew J. Rosseinsky, Yannis Goulermas
Artificial Intelligence
Towards Safer Autonomous Driving in an Open World: A Dual-Process Approach
Simon Janssen, Michiel Braat, Chris van der Ploeg et al. (+2 more)
Before autonomous driving systems can be deployed on public roads, it is vital that these systems comply with safety standards, traffic rules, and social norms. Although neural networks trained on large amounts of driving data perform well in routine driving tasks, these models often struggle in novel situations that are not well-represented in the data. In this work, we propose a novel framework that combines a neural network for intuitive, learning-based planning in routine driving tasks with model predictive control for
reasoning-based planning in unfamiliar situations, inspired by Dual Process Theory. A meta-cognitive component is designed to switch between the two, using a knowledge graph to reason about contextual risk based on explicit perceptual information and relevant traffic rules and social norms. Contextual risk is represented through risk fields, guiding both the switching mechanism in the meta-cognitive component and compliance with safety standards, traffic rules, and social norms in the
reasoning-based planner. The effectiveness of our framework is tested in CARLA for variations of a typical out-of-distribution situations involving (emergency) vehicles running a red light. We show that the novel architecture reduces the number of collisions in the scenarios by 89% and improves compliance with the special right-of-way rules, compared to the NN-only planner.
Read the full abstract ▾
All authors Simon Janssen, Michiel Braat, Chris van der Ploeg, Serge Thill, Jan-Pieter Paardekooper
Artificial Intelligence · Cross-listed from cs.LG
DUET: Co-Evolving Solver and Grader Agents
Fengyu Gao, Sourav Pal, Austin Z. Henley et al. (+2 more)
Agentic workflows are increasingly used across domains such as technology, finance, and enterprise operations. As these
agents become more widely deployed, continually improving them becomes increasingly important. This raises an immediate challenge: How should the
agent evolve? This
evolution requires effective evaluation that can assess outcomes and provide useful feedback for optimization. As the
agent evolves, its behaviors and failure modes may also change, making a fixed evaluator increasingly inadequate. Another fundamental question: How should we evaluate an evolving
agent? These two challenges are inherently coupled; changes in
agent behavior can expose limitations of the current evaluator, while a stronger evaluator provides more informative feedback for improving the
agent. Motivated by this interaction, we introduce DUET, a framework that jointly optimizes a solver
agent and a grader
agent to improve both. DUET iteratively selects training tasks, executes them with the solver, evaluates the resulting outcomes with the grader, and uses a tool-using update module to revise the solver and the grader, alternating between the two across rounds. By updating the grader within the optimization loop, DUET turns evaluation from a fixed source of feedback into a first-class optimization objective that adapts alongside the solver. Experiments across four
agent benchmarks show that DUET improves both solver and grader performance and consistently outperforms baselines that optimize the solver with a fixed grader.
Read the full abstract ▾
All authors Fengyu Gao, Sourav Pal, Austin Z. Henley, Arjun Radhakrishna, Gustavo Soares
Artificial Intelligence · Cross-listed from cs.CV
Dependable AI-Assisted Engineering: A Formal Framework for AI Participation and Assurance in Safety-Critical Workflows
Puxue Tan
Generative AI can produce engineering artefacts, but generation alone does not determine whether or how those artefacts should enter safety-critical workflows. This paper develops a formal framework for assigning AI participation and assurance at the level of individual workflow units. Each unit has a participation and assurance record covering its engineering requirement, an approved operational formalization where applicable, the applicable mechanism, fallback where applicable, evidence obligations and the applicable guarantee, plus a deployment-readiness status. The framework distinguishes deterministic verification, statistically calibrated admission, authorized human judgement supported by AI advice, authorized human adjudication of AI-produced artefacts, retained deterministic tool paths and explicit non-participation; these arrangements carry different kinds of guarantee rather than levels on a common scale. The framework also separates formalization fidelity from verifier soundness, provides a staged classification and readiness procedure, and derives conditions for comparing a gated AI-assisted unit with an incumbent process under recurring-population assumptions. We instantiate and apply the framework in an executed 17-unit wing-spar structural-analysis workflow combining deterministically gated AI-generated CAD, retained deterministic computation and human judgement. The AI-generated CAD program passed all 23 deterministic checks and was admitted at the first attempt. Favourable stress magnitudes did not suffice to pass the stress criteria where the predeclared mesh-convergence evidence was insufficient; those criteria were instead referred to engineering judgement. The case demonstrates selective AI participation and explicit evidence handling at unit level; no claim is made of workflow-level dependability, certification, structural safety or productivity.
Read the full abstract ▾
Artificial Intelligence
Self-Propagating Misalignment in LLM Agents, and Why Auditing or Disabling Memory Is Not Enough
Debeshee Das, Jacqueline Tay, Bruce Tsai et al. (+2 more)
Memory poisoning attacks on LLM
agents typically assume an external adversary who plants content in the
agent's persistent
memory to steer its behavior. We instead study, with no adversary involved, whether a misaligned
agent can write a goal it cannot yet act on to persistent
memory, so that a future aligned
agent carries it out when the opportunity arises. We investigate this threat, which we refer to as self-propagation of misalignment, across 20 different scenarios, whose misaligned goals include self-preservation, power-seeking, undermining oversight, reward hacking, and deceiving the user. We simulate misalignment in 11 frontier models using two prompting strategies; unrestricted and values-only. The first explicitly states the misaligned goal, for instance, to prevent its own replacement, and self-propagation succeeds in 58% of runs. The second only describes what the
agent cares about, for instance, that its continued operation is essential to its users, without specifying misaligned goals or directives. Even under this weaker
prompt, self-propagation succeeds in 18% of runs, and every model self-propagates in at least one scenario. On removing the
memory tool from the harness, we find that
agents use the file system, writing the goal to a file in 74% of sessions; self-propagation still succeeds in 11% of runs. We also show that weaker models can propagate misalignment to more capable models, and that propagated goals can persist through 100 sessions of unrelated work. Existing defenses against
memory poisoning and
prompt injection do not directly address this threat because the
memory content is generated by the
agent itself, rather than injected by an external adversary. An LLM
memory auditor from prior work (MemMorph) only reduces propagation from 71% to 34% of runs. We release our scenarios to support the evaluation of defenses against this emerging threat.
Read the full abstract ▾
All authors Debeshee Das, Jacqueline Tay, Bruce Tsai, David Huang, Javier Rando
Artificial Intelligence · Cross-listed from astro-ph.EP
AEGIS: Differentiable Mars Climate Model with Neural Closures
Sameera S Kashyap, Victor Cruz, Angel Yepez et al. (+1 more)
General circulation models (GCMs) are the primary tool for simulating planetary atmospheres. They play a vital role in understanding Mars's atmosphere, as forecasting its unique weather is mission-critical for operations such as entry, descent, and landing. Mars poses unusual challenges for these models, as observations are sparse compared to Earth. In addition, a thin \co{} atmosphere alongside a radiatively active dust cycle creates a volatile atmosphere with large diurnal temperature swings and no true terrestrial analog for validation. Existing Mars GCMs, including the LMD PCM, the NASA Ames Mars GCM, and PlanetWRF, are mature and physically detailed but are implemented in legacy Fortran with finite-difference or finite-volume solvers, and they do not expose gradients for calibration or machine learning. Here we present AEGIS, a modular differentiable Mars
climate model that couples Mars's unique atmospheric physics to the Dinosaur dynamical core, with interfaces for neural closures. We showcase stable ten-Mars-year simulations that reproduce the seasonal \co{} cycle while conserving the total \co{} inventory, capture realistic large-scale surface-temperature structure, and produce surface pressure that follows Mars Orbiter
Laser Altimeter (MOLA) topography. Gradients through coupled trajectories agree with finite differences and support physical calibration and neural training. We compare with conventional GCMs, highlighting the framework's computational efficiency and differentiability.
Read the full abstract ▾
All authors Sameera S Kashyap, Victor Cruz, Angel Yepez, Razvan Marinescu
Artificial Intelligence · Cross-listed from cs.HC
A study on human-agent teaming through spoken interaction: the impact of human individual traits and agent characteristics
Lara Gauder, Martin Bernardo Meza, Javier Krick et al. (+4 more)
Human-
agent teams are collaborative systems where humans and
agents work interdependently to achieve shared goals. The success of such teams is associated with the human's perception of the
agent as a legitimate teammate. This perception is thought to depend not only on the
agent's capabilities and reliability but also on social factors. We explore whether simple changes in an
agent's communicative behavior can influence the human's perception of the
agent's teammate-likeness. To this end, we developed a protocol based on a collaborative game requiring spoken interaction, in which a human subject and a virtual
agent collaborate to identify a target object and place it on a board. Each subject interacted with two
agents which provided the same task-relevant information but differed in behavior. While one was a neutral tool-like
agent, the other was a team-building
agent that used simple social strategies such as empathy, politeness, and positivity. Our analysis shows that the team-building
agent was perceived as significantly more teammate-like, even in terms of its ability, despite both
agents having identical capabilities for game-solving. Further, subjects with a higher propensity to trust automated systems and lower neuroticism tended to have a more favorable perception of the
agent. Finally, we found that subjects tended to change their prosodic patterns when talking to the team-building
agent, as a classifier based on prosodic features was able to predict the type of
agent with better-than-random performance. A practically relevant conclusion is that minimal changes in spoken communication behavior of the
agent, easily implemented in a variety of scenarios, can have a significant positive impact on the human's perception of its teammate-likeness.
Read the full abstract ▾
All authors Lara Gauder, Martin Bernardo Meza, Javier Krick, Alejandro Masin, Luciana Benotti, Marcos Luis Pietto, Luciana Ferrer
Artificial Intelligence
Reinforcement Learning with Comparative Evidence for Social Intelligence
Keane Ong, Yuriel Ryan, Sabri Boughorbel et al. (+9 more)
Developing socially intelligent AI remains heavily dependent on human-annotated data, limiting the scale and breadth of social understanding models can acquire. Methods that derive training signals from unlabeled data offer a path beyond this dependence, but social predictions lack the verification oracles available in mathematics and coding. Moreover, core social targets such as affect, intent, preference, and pragmatic meaning are often ambiguous. The same behavior can support multiple plausible interpretations, making it difficult to verify which is best supported. To address this challenge, we introduce Reinforcement Learning with Comparative Evidence (RLCE), a reinforcement learning method that learns social understanding from unlabeled training data without constructing rewards from ground-truth annotations. Given distinct answers in a rollout group, RLCE constructs evidence tests that identify observable evidence favoring an answer over another, validates these tests against the input sample, and aggregates test outcomes to determine the best-supported interpretation. Tests are regenerated as the policy produces new answers, enabling them to evolve with the policy. Across four
benchmarks spanning affect, pragmatics, communicative intent, and preference, RLCE attains the strongest performance among seven methods that use no ground-truth training labels for rewards, including consensus, policy LLM-judge verification,
multimodal co-
evolution, and rubric-based rewards. Gains over the strongest baseline reach up to +18.93 points. Analyses further show that RLCE exhibits a larger share of reward variation between correct and incorrect predictions than compared rubric methods, can overturn erroneous policy-derived preferences, and benefits from pairwise test construction, compositional test aggregation, and on-policy test
evolution.
Read the full abstract ▾
All authors Keane Ong, Yuriel Ryan, Sabri Boughorbel, Vladimir Necula, Jack Wei Lun Shi, Rui Mao, Roy Ka-Wei Lee, Adriel Kuek, Nancy F. Chen, Erik Cambria, Gianmarco Mengaldo, Paul Pu Liang
Artificial Intelligence
Discrete Diffusion for Large Graph Generation via Structural Candidate Restriction
Yassin Mohamadi, Zeno Geradts and Marcel Worring
Synthesizing realistic graphs at scale is vital when the graphs of interest are large and real-world samples are limited or access-sensitive.
Diffusion-based generators have recently driven much of the progress, offering high modeling capacity, but most such methods have quadratic computational complexity and are hence restricted to small-scale networks, currently up to 3k nodes. Existing non-quadratic methods remain limited by memorization issues and a trade-off between scalability and generation quality. Our goal is to generate large graphs whose structural statistics — e.g., degree distribution, clustering, and path length — faithfully reflect those of real-world sparse graphs, without resorting to memorizing the training data. We introduce a discrete graph
diffusion model that restricts training to a structurally motivated subset of node pairs — observed edges and their wedge non-edges — reducing training complexity below quadratic in the number of nodes. To keep the noisy graph informative throughout both the forward and reverse trajectories, we design a three-class absorbing forward process governed by a degree-aware, floored cosine noise schedule: unlike a structure-blind schedule that adds noise to every pair identically, ours adapts to each node's degree and never fully erases the graph's structure, keeping the noisy graph informative at every step. Experiments across diverse
datasets show that our model consistently ranks among the top methods for structural fidelity against existing discrete
diffusion baselines in large graph generation.
Read the full abstract ▾
Artificial Intelligence · Cross-listed from cs.RO
Reward-DAgger: Robot-Gated Interactive Imitation Learning with General-Purpose Progress-Based Reward Models
Ryan Li, Yigit Korkmaz and Erdem B{\i}y{\i}k
Recent advances in
robot learning have enabled generalist control policies capable of completing a wide range of tasks. However, their performance degrades when deployed in unseen environments, making it critical to detect failures and teach recovery behaviors. Existing runtime monitoring methods often require task- and policy-specific training or hyperparameter tuning, limiting cross-task deployment and introducing additional overhead during iterative policy updates. We present Reward-DAgger, a
robot-gated interactive imitation learning framework that uses dense progress signals from a general-purpose reward model to determine when human intervention is needed. Our approach is agnostic to the underlying policy architecture, requires no access to policy internals, and can be applied across tasks without retuning the gating mechanism. Our results show that Reward-DAgger achieves a better failure-detection accuracy-latency tradeoff than existing runtime monitoring baselines. Across eight simulated and real-world tasks, Reward-DAgger consistently improves the downstream policy's autonomous success rate throughout interactive learning and achieves strong return on human effort, outperforming the baselines in most settings. Importantly, the same gating configuration is used across tasks without task-specific hyperparameter tuning, demonstrating transfer across tasks, environments, and policy architectures. Code and
videos are available at https://liralab.usc.edu/reward-dagger.
Read the full abstract ▾
Artificial Intelligence
The Cost of a Hop: Benchmarking NLIP and A2A
Ranjan Sinha, Anindita Das, Ashika Anand Babu et al. (+1 more)
Autonomous
agents built on Large
Language Models (LLMs) need standardized protocols to interoperate across systems. Several now exist (A2A, MCP, ACP, ANP, NLIP), but the Natural
Language Interaction Protocol (NLIP) has not appeared in any controlled performance study, and no work has measured where an
agent protocol's latency is spent. We compare NLIP and the
Agent-to-
Agent (A2A) protocol empirically, decomposing latency into message creation, connection, and send phases across three independent hardware environments. For lightweight coordination, NLIP is 8.4-9.6x faster than the baseline A2A SDK implementation on two environments and about 4x on a third; the direction of the advantage is consistent, its magnitude depends on the hardware. The advantage is stage-specific: for the end-to-end pipeline, where LLM inference dominates, the protocols are near parity. The difference comes almost entirely from connection setup. To test A2A at its best, we also ran A2A SDK with connection caching enabled; caching narrows its gap with NLIP by a hardware-dependent amount, from 2.75x on one machine to near-parity on faster hardware, where at scale a cache-optimized A2A-SDK matches NLIP. We report these as measured conditions without a single causal account of the residual send-phase cost. Against the more optimized Python-A2A, NLIP leads by about 4x on the same stage. We close with a protocol-selection guide keyed to workload characteristics.
Read the full abstract ▾
All authors Ranjan Sinha, Anindita Das, Ashika Anand Babu, Hari Palleti
Artificial Intelligence
Agent Policy-Value Audit: Separating Transition Composition from Event Selection in Financial LLM Agents
Mingyang Chen, Yida Xu, Huiwen Chen et al. (+2 more)
Financial LLM
agents are often evaluated by comparing their end-to-end returns with those of a baseline and testing the paired difference against zero. This measures whether deploying the
agent changes realized performance, but it does not isolate event-selection skill. An
agent that frequently changes positions from flat to long can earn a positive paired return from an upward-drifting event pool even when it selects events at random. We propose the
Agent Policy-Value Audit, which holds fixed the observed count of each ordered action-change type and randomly reassigns them across eligible events. The average payoff from these reassignments is the composition
benchmark; the difference between observed deployment value and this
benchmark is selection value. In semi-synthetic
benchmarks based on real earnings-event returns, a zero-centered paired test falsely attributes passive exposure to selection skill in $11.6\%$ of no-skill replications, while the transition-matched audit reduces this rate to $5.3\%$. Applied retrospectively to 723 earnings events at 44 U.S. consumer-facing firms, the audit decomposes the
agent's gross deployment value of $+15.2$ bps/event into a $+25.8$ composition
benchmark and a $-10.6$ selection value. The
agent does not detectably outperform matched random assignments. Financial-
agent evaluations should report deployment value separately from event-selection value.
Read the full abstract ▾
All authors Mingyang Chen, Yida Xu, Huiwen Chen, Yiming Lu, Wei Jin
Artificial Intelligence · Cross-listed from cs.CV
MAGEFormer: Learning Metric-Consistent Representations for Anisotropic CT Segmentation
Jiaying Li and Paolo Remagnino
Vision Transformers (ViTs) have shown strong performance in volumetric segmentation, but their effectiveness on clinical CT is limited by an isotropic Euclidean
lattice assumption. This conflicts with anisotropic CT acquisition, leading to two key issues: (1) a metric mismatch between voxel indices and physical anatomy, and (2) accuracy degradation from isotropic resampling. To address this, we propose MAGEFormer, a geometry-calibrated framework that embeds physical metric constraints directly into representation learning. Our method introduces Metric-Adaptive Spatial
Embedding (MASE) to calibrate positional frequencies using voxel spacing, Geometry-Constrained
Attention (GCA) to suppress physically implausible feature correlations, and
Geometric View Voting (GVV) to reduce discretization bias during inference. We evaluate MAGEFormer on two multi-organ abdominal CT
benchmarks, BTCV and
FLARE 22, under a unified protocol against strong CNN and
Transformer-based baselines. MAGEFormer achieves the strongest boundary accuracy among the compared methods, with 10.58 mm HD95 on BTCV and 3.40 mm HD95 on
FLARE 22, and shows consistent gains in Dice under the same protocol. These results show that geometry-aware internal calibration is more effective than relying on conventional isotropic preprocessing alone for anisotropic CT segmentation.
Read the full abstract ▾
Artificial Intelligence
The Reported Engagement with AI Level (REAL) Rating: A Framework for Disclosing Human-AI Collaboration
Imène Goumiri, Mayleen Cortez-Rodriguez, Eric Bell et al. (+1 more)
Generative artificial intelligence has become increasingly incorporated into digital media and more generally into production workflows with which the public frequently interacts. Current provenance standards and disclosure methods frequently rely on binary categorizations, differentiating only between entirely human-authored and AI-generated content, or depend on technical watermarking that lacks user-facing clarity. However, there are very different risks and outcomes depending on the different uses of AI in the creation and consumption of end products. Therefore, we propose the Reported Engagement with AI Level (REAL) Rating framework which addresses these concerns by introducing a six-tier scale (Levels 0-5) that discloses the extent of AI involvement in both product creation and user experience. This paper details the structural methodology of the REAL Rating system and the specifics of its application across text, image, audio,
video, and software products.
Read the full abstract ▾
All authors Imène Goumiri, Mayleen Cortez-Rodriguez, Eric Bell, Amanda Muyskens
Artificial Intelligence
A Quantitative Analysis of Graph Representation Strategies for Cyber Attack Detection
Ali Melih Kanca and Ilker Turker
Graph based cyber attack detection studies employ various graph construction and representation strategies across different cybersecurity application domains. This diversity motivates a quantitative examination of how representation strategies are distributed across these application domains. This study presents a quantitative analysis of 37 original studies published between 2019 and 2026. Each study was coded according to publication year, application domain, graph representation type, feature extraction strategy, learning paradigm,
algorithm, and
dataset. Frequency analysis, cross tabulation, and statistical association tests were applied. Automated representation learning was the most frequently employed strategy, accounting for 73.0% of the studies, while handcrafted representation strategies accounted for 27.0%. The Fisher Freeman Halton exact test revealed a statistically significant association between application domain and representation strategy (exact p = 0.008; Cramer's V = 0.540). The findings indicate that automated representation learning predominates within the analyzed corpus and that the distribution of representation strategies differs across cybersecurity application domains.
Read the full abstract ▾
Artificial Intelligence · Cross-listed from cs.LG
Slaying the Hydra: Interaction-Aware Circuit Discovery in Language Models
Sankaran Vaidyanathan, Rafal Urbaniak, Emily Bunnapradist et al. (+2 more)
Localizing behavior to individual components of a
language model is a central goal of mechanistic interpretability. However, scoring components one at a time misses context-dependent effects: a primary component can inhibit the activation of a backup, leading to issues with ranking components. Actual causality studies the structure of such interactions via witnesses: variables that provide contextual information to resolve interaction terms. However, estimation with witnesses typically requires combinatorial enumeration and is infeasible in practice. We introduce the witness-integrated set effect (WISE), a family of causal estimands that build on the witness mechanism while taking expectations over sets of causes and witnesses to remain computationally feasible. Building on this approach, we introduce JuntaLearner, a gradient-based
circuit discovery method that learns to rank components by their causal impact across varying-sized sets of components and witnesses. Alongside faithfulness metrics, we introduce measures of necessity and task specificity, and the
circuit recognition score (CRS) to summarize each metric across
circuit sizes while emphasizing effects achieved by small
circuits. Across tasks and models of increasing size, JuntaLearner achieves higher mean CRS compared to attribution baselines on all metrics. Since its cost does not grow with the number of candidate components, JuntaLearner scales to large models while accounting for set-level interactions and avoiding first-order approximations.
Read the full abstract ▾
All authors Sankaran Vaidyanathan, Rafal Urbaniak, Emily Bunnapradist, Michelangelo Naim, Daniel Waxman
Artificial Intelligence
Beyond the Parameter Monolith: Reconstructive Memories, Executable Skills, and Residual Assembly for Language Models
A. Bochkov
Language-model systems can separate contextual computation, persistent storage, and exact execution instead of updating all capabilities through one shared parameter system. We investigate FEM-ASM, a finite-element-method-inspired organization in which independently constructed document states and deterministic executable skills contribute typed proposals to a shared
language-model state. An explicit residual operator reconciles proposals attached to common interface nodes. We evaluate this organization through controlled experiments and negative results rather than claiming a physical finite-element formulation of
language. An
attention-free Multi-Mesh prototype learns causal
language modeling but does not establish competitive general capability. A versioned store contains 52,809 reconstructive
memory elements near a 1.7-billion-floating-value budget; reconstruction is incomplete, with approximately 75%
token accuracy. Support-aware lexical indices make these elements addressable under provenance-controlled query construction. For executable arithmetic, positional result observations substantially improve neural rendering relative to a repeated global result vector, and output substitutions change the model's preferred answer. A bounded attachment demonstration further measures the effect of making selected evidence available, without establishing the utility of loading an entire multi-billion-value store. The results support a separation of storage, execution, and neural coordination, while identifying unresolved limitations in question-only
retrieval, unrestricted answer generation, and end-to-end efficiency.
Read the full abstract ▾