Scaffolding GenAI Use in Ethical Essay Writing: A Three-Year Experience Report in Introductory Computing
Asma Shakil, Paul Denny, Jun Seo, Gauravkumar Hirani, Sean Longyu Ma, Claus Brabrand
AI 速览：该经验报告基于三年六次调查与成绩关联分析，探讨在一年级计算机系统课程的伦理短文作业中，如何通过支架式负责任使用流程引导学生使用生成式AI。
As society, ethics, and professional practice become increasingly integrated across the computer science curriculum, technical courses need practical ways to support ethical reasoning. Ethical essay assignments offer one approach, but little is known about how students use GenAI when writing such essays. This experience report examines GenAI use across three years of ethical essay assignments in a large first-year computer systems course. Across six post-essay surveys from 2024 to 2026, we analyzed 1,894 responses, linked survey data with essay grades in 2025 and 2026, and introduced a scaffolded responsible-use workflow in 2026. Under voluntary-use policies, AI adoption remained steady at around two-thirds of students, rising above 90% when use was scaffolded. Students reported mainly using AI for help with grammar and style, while tool use diversified beyond ChatGPT in 2026. Grade-linked analyses showed that AI use did not independently predict essay quality, while marker evaluations indicated high levels of critical engagement when use was scaffolded. Overall, GenAI's pedagogical value appears to depend on how assessment design guides students to evaluate and critically engage with AI-generated suggestions.
Angela Sara Cacciapuoti, Jessica Illiano, Caterina De Risi, Marcello Caleffi
The Quantum Internet is commonly envisioned as a network that transports quantum states and distributes entanglement as the key resource underpinning distributed quantum applications. This view is necessary but incomplete if quantum phenomena are confined to what the network carries or manages, while network functionalities remain grounded predominantly in classical abstractions. This Perspective advances a broader thesis: quantum phenomena can also participate directly in the realization of network functionalities. We first establish the architectural baseline by introducing entanglement as a stateful network resource whose operational condition is non-locally distributed across remote nodes, evolves through manipulation or consumption, and degrades over time through decoherence. Its evolving state must therefore remain visible to the functionalities acting on it. We then move beyond resource visibility and use two mechanisms as proof points of a broader quantum-native functional design space: entanglement generation during propagation through coherent spatial-path superposition, and operational connectivity induced by shared entanglement that can depart from the physical graph. These examples show that paths, neighbors, or participating nodes may become operationally relevant in ways determined by quantum coherence and shared entanglement, rather than by physical connectivity alone. We identify Quantum Addressing as a transversal abstraction capable of representing such relations while complementing, rather than replacing, classical addressing and signaling. The resulting perspective broadens the Quantum Internet from transporting quantum information and distributing entanglement to exploiting quantum phenomena in network functioning, opening a coupled frontier spanning routing, forwarding, control, and functionality placement.
Recent advances in speech generation have spurred the development of speech editing techniques. However, achieving high-quality speech editing remains challenging due to two key limitations. First, the major hurdle is the lack of large-scale, high-quality speech editing datasets. Constructing such datasets is costly, while open-source, scalable data pipelines remain limited. Second, the learning dynamics of speech editing models remain poorly understood. In particular, how models acquire editing capabilities and the relationship between generation and editing remain unclear. To address these challenges, we introduce VoxCanvas-41M, a large-scale bilingual dataset comprising approximately 41 million speech editing pairs. VoxCanvas-41M covers two major categories: content editing, including single-span, multi-span, and location-reference editing, and paralinguistic editing, covering the insertion and deletion of 39 classes of nonverbal events such as laughter, sighs, and coughs. Building on VoxCanvas-41M, we train VoxCanvas-Edit, an instruction-based speech editing model, demonstrating the effectiveness of our dataset in improving editing performance. Furthermore, we systematically investigate the learning dynamics of speech editing, examining how models learn different editing operations and how data scale, task interactions, and training strategies affect editing performance. The dataset, pipeline, and demos are publicly available.
Michael Brown, Maryam Rahimimovassagh, Mehdi Yazdani-Jahromi, Sanjeda Sara Jennifer, Ivan Garibay
Diffusion-based molecular docking requires learning a corrective score field across a broad corruption distribution, yet individual pose errors can differ substantially in their structural consequences. This creates a post-training conflict between preserving broad score-field coverage and concentrating supervision on corrections that appear more relevant to docking geometry. Using DiffDock, a fixed-budget study compares broad continuation with torsional supervision focused on Cartesian consequence or conditional target consistency, and with a broad/focused mixture. Evaluation separates fixed score fidelity, generated-pose quality, best-of-10 sampling capacity, and confidence-selected Top-1 docking performance. At <5 Angstrom, broad continuation had the highest mean Top-1 success among trained policies, and focused or mixed supervision showed no consistent advantage under the tested budget. The starting model had better fixed score fidelity, whereas the evaluated broad-continuation model had higher mean Top-1 success with the same aggregate best-of-10 recovery. Focusing supervision on structurally consequential or more consistently specified corrections did not consistently improve returned-pose success over broad continuation.
Teng Pan, Zixuan Wang, Yuchen Yan, Qifei Wu, Yao Chen, Yingqi Qu, Ruiqing Zhang, Yongliang Shen
Large language models have made rapid progress in agentic reasoning and tool use, raising expectations for agents that can take over realistic productivity work from humans. However, current agents still struggle to finish such tasks, where success requires following long-horizon professional workflows, understanding heterogeneous file contexts, and producing usable deliverables under fine-grained requirements. A key bottleneck is the lack of scalable post-training data that reflects real work intent, authentic file environments, and deliverable-level evaluation. To close this gap, we introduce Agent-R$^2$, a $\textbf{r}$eal-source data synthesis framework that scales training data for $\textbf{r}$ealistic productivity work. Instead of relying on manually specified workflows or LLM-generated files, Agent-R$^2$ grounds synthesis throughout the pipeline in real occupational seeds, corpus-derived workflow evidence, and authentic web-sourced files. From 10k web-corpus candidates, Agent-R$^2$ distills 4k task instances that pair real multi-format attachments with fine-grained, deliverable-level rubrics. On this data, we train Agent-R$^2$-4B via supervised fine-tuning and rubric-based reinforcement learning, improving both end-to-end productivity performance and file-grounded reasoning. Trained on only 4k instances, Agent-R$^2$-4B surpasses Qwen3-235B-A22B-Thinking-2507 on both GDPval (51.0$\%$ vs. 48.0$\%$) and OfficeQA (14.8$\%$ vs. 13.8$\%$). Further analysis shows that Agent-R$^2$ mitigates shortcut behavior, where agents produce superficially plausible deliverables before sufficiently grounding, constructing, or verifying them, while promoting sustained exploration.
Dynamic flexible assembly flow shop scheduling with multi-product delivery, often with kitting supply and machine-level flexibility, can be found in many applications. It is difficult to design rules manually for it. Genetic programming (GP) has been used to automatically generate dispatching rules, enabling responsive scheduling while meeting high responsiveness demands. However, GP methods are dependent on fixed terminal sets and have poor interpretability. In this work, we develop an evolving dispatching rules framework (EvoDR) that leverages the semantic understanding and generation capabilities of large language models to achieve cross-domain integration of rule design and scheduling knowledge. In our proposed framework, multi-stage assembly supply decisions are modeled as priority sorting of directed edges based on heterogeneous graphs. A dual-expert co-evolution mechanism is proposed, which allows LLM-A to generate code and LLM-S conduct scheduling analysis and reflection. Guided by improvements in hybrid evaluation, adaptive rules that fit dynamic features are continuously evolved. Experimental results show that EvoDR can achieve lower average tardiness than other automatically designed state-of-the-art approaches. It also outperforms expert-designed competitors across 480 instances over 24 distinct scenarios with varying resource configurations and disturbance levels, and EvoDR demonstrates robust adaptability across diverse operating conditions.
Generative creative interfaces routinely report the user's position within a space of possibilities. Such displays imply that the quantities shown were measured from the artefact; in a growing class of systems they are instead predicted by a language model describing its own output, and the two are indistinguishable on screen. Most creative domains lack an agreed measurement against which such a self-report can be checked; audio admits one. In a measurement study of a working LLM-driven timbre interface, what a user can cause proves separable from what the interface reports. The first is exactly characterisable: the reachable set is a zonotope in parameter space whose dimension increases by one per sound placed, so the space the model may act within is demonstrably the user's own. The second is inaccurate. The displayed position differs from measurement by approximately three times the width of that region, and the error has two independent sources: model-predicted coordinates, whose accuracy varies twofold between the two models evaluated, and the interpolation step, which arises from combining synthesiser parameters and is unaffected by model quality. End to end, 27 of 36 refinement instructions produced no perceptible change, and under the weaker model the facility performed no action while reporting convergence. Effective agency (what a user can cause) and epistemic agency (whether they can know it) are therefore separable and independently engineered: structurally enforced constraints survive model failure, whereas guarantees resting on the model's own report do not. The failure mode that remains is silence presented as success. We close with a short protocol for building and auditing such displays.
Large-scale identity-labeled face datasets are essential for training modern face recognition models, yet collecting and annotating such data is costly and raises privacy and ethical concerns. Existing synthetic face generation methods typically derive identities from predefined class labels, unconditional image generation, or predefined latent sampling strategies, without explicitly modeling the underlying distribution of facial identities. In this paper, we propose IDPrior, a two-stage framework that decouples identity generation from image synthesis. We first learn a diffusion prior in the embedding space of a pretrained face recognition model to model the distribution of facial identities and generate novel identity embeddings. These embeddings are then used to condition a pretrained identity-conditioned diffusion model (DM), enabling the synthesis of multiple images of the same identity. This modular design provides a scalable approach to identity generation while remaining compatible with existing and future identity-conditioned DMs. Experiments indicate that IDPrior generates diverse and well-separated identities while maintaining a closer correspondence to the distribution of authentic identities. Face recognition models trained on datasets generated using IDPrior-generated identities achieve competitive or improved performance compared with existing synthetic-data-based methods across multiple face verification benchmarks. Code and generated datasets are publicly available at https://github.com/fdbtrs/IDperturb.
Goal interpretation is to translate a natural language instruction into a formal symbolic goal that a planner can execute. It is critical for embodied agents because any error here propagates to every downstream planning and action step. Doing it well requires both generating suitable candidates and reliably selecting among them. We present a two-stage generate-then-rank approach to Goal Interpretation in the Embodied Agent Interface (EAI). First, supervised fine-tuning adapts a Large Language Model (LLM) to EAI's output convention. Second, Candidate-Output Ranking from Errors (CORE) learns preferences over complete outputs from the model's own generated alternatives. We conducted experiments on the household scenes of the Embodied Agent Interface Benchmark. Multiple open source LLMs including DeepSeek-V3 chat, DeepSeek-Reasoner, and Qwen2.5-7B and Qwen2.5-72B were evaluated. The proposed method outperformed the baseline models including LLMs with simple prompting, chain-of-thought, and knowledge graph based approaches.
Linear Complementarity Problems (LCPs) with sufficient matrices form a well-studied class of problems, dating back from the foundational works of Cottle and Dantzig (1968). LCPs capture convex quadratic programs, and admit reductions from problems such as discounted stochastic games or mean-payoff games. While the original approach for solving these problems was based on pivoting methods, in their 1991 book, Kojima, Megiddo, Noma, and Yoshise introduced an interior point method (IPM) with running time that is polynomial in the input size, but which depends linearly on a certain parameter called the handicap number of the matrix. Despite the ensuing rich literature on sufficient LCP IPMs, the linear dependence on the handicap number has not been improved since then. We present an IPM whose running time depends on the square root of the handicap number. The algorithm can decrease the duality gap by a constant factor in $O\left(\min\left\{\sqrt{n} (1+κ),n^{9/8}\sqrt{1+κ}\right\}\right)$ iterations. Curiously, this can already be achieved using standard Newton updates. We analyze the IPM in a norm defined by a local weighting, where centralizing Newton steps quadratically converge to the central path, without any dependence on the handicap number. The weights arise from a characterization by Rump (Linear Algebra and Its Applications, 1997, 1998, 2003) of the sign-real spectral radius and its relation to sufficiency; another ingredient of our proof is a new operator norm characterization of the handicap number. Moreover, the connection to Rump's work also enables us to show coNP-completeness of testing sufficiency, and approximation hardness of the handicap number.
Lucile Favero, Tanja Käser, Juan Antonio Pérez-Ortiz, Nuria Oliver
Supporting students' argumentative writing with AI requires connecting reflection to their revision processes. We present Maike, a Socratic chatbot supporting writing through argument diagnosis, targeted questioning, and a student model across Think and Write modes. In a classroom study, 65 secondary-school students used Maike or a control chatbot sharing the interface, language model, and Think/Write modes. Approximately 40% of students improved their essay scores. However, Maike users engaged more in idea-level dialogue (56% vs. 44%) and moved iteratively between discussion and editing. Transitions from ideas discussed in dialogue to revisions were positively associated with score gains. These findings suggest a design target for AI writing support: coordinating reflection with revision. We contribute an architecture for supporting argumentative reasoning without generating essay text and an approach linking interaction patterns to writing processes and outcomes. As AI becomes embedded in classrooms, understanding these connections is essential to assessing (and designing for) its educational value.
Nirmit Joshi, Roey Magen, Nathan Srebro, Nikolaos Tsilivis, Gal Vardi
Chain-of-thought (CoT) supervision makes computationally hard learning problems tractable by exposing intermediate steps that are easy to learn via next-token prediction. However, in practice, CoT datasets draw on heterogeneous sources with different reasoning styles, and substantial effort goes into curating supervision suitable for next-token training. What role does this curation play in enabling the computational advantage of CoT supervision? In this work, we take a first step toward understanding this question from a learning-theoretic perspective. We introduce several models of $heterogeneous$ $CoT$ $supervision$ from multiple sources, each providing correct traces with the same final answer but different intermediate steps. Assuming that fitting a next-token predictor to $homogeneous$ traces from any single source is computationally tractable, we ask when this suffices for efficient learning from heterogeneous supervision. We show $cryptographic$ $hardness$ of learning from $passively$ collected traces from multiple sources, even though learning from homogeneous CoTs is easy. This hardness can stem either from $hard$-$to$-$learn$ $contribution$ $biases$ determining which sources contribute traces on different parts of the domain, or from receiving an $unfiltered$ heterogeneous mixture (though a filtered subset of homogeneous traces is sufficient for efficient learning). On the positive side, inspired by iterative and adaptive fine-tuning of language models, we give a computationally efficient learning algorithm that $actively$ and $adaptively$ collects CoTs from multiple sources with a per-source CoT budget completely independent of the target error $\varepsilon$. We argue that our results illuminate data collection and curation practices central to making CoT supervision effective for learning.
Expert pruning compresses Mixture-of-Experts (MoE) language models by removing the experts that a calibration set marks as unimportant. Studies of high-resource languages largely find that English calibration transfers well to other languages. We test this for Lithuanian: we pruned four MoE models with REAP, removing 25% or 50% of the experts with English, Lithuanian or mixed calibration, and evaluated both languages on reading comprehension, translation and language modelling. English calibration does not transfer to Lithuanian: at 25% pruning, English comprehension fell by a median of 1.7 points but Lithuanian by 21.4. The transfer is asymmetric. Lithuanian calibration removed most of the Lithuanian loss and never significantly hurt English comprehension, which was on average 2.5 points better than under English calibration; only English language modeling got slightly worse. Routing suggests why: in the last layers, Lithuanian sends 55-72% of its expert selections to just eight experts, against 19-39% for English, and English calibration removes some of these experts. Compressed models should be calibrated, and evaluated, in every language they must serve.
Central to human-aligned AI is understanding the benefits of human-elicited labels over synthetic alternatives. While human soft-labels improve calibration by capturing uncertainty, prior studies conflate these benefits with the implicit correction of mislabeled data (mode shifts), obscuring true effects of soft-labels. We present a controlled audit of soft-label learning across MNIST and a synthetic variant, re-annotating subsets to extract human uncertainty. By decoupling soft-label supervision from underlying label mode shifts, we show that while human soft-labels do provide accuracy gains, their larger value lies in acting as a regularizer that improves model calibration on difficult samples and promotes stable convergence across training runs. Dataset cartography reveals models trained on human soft-labels mirror human uncertainty, whereas those trained on synthetic labels fail to align with humans. Broadly, this work provides a diagnostic testbed for human-AI uncertainty alignment.
Speech recognition still works poorly for many people whose speech is affected by conditions such as ALS, Parkinson's disease, cerebral palsy, Down syndrome or stroke. We took part in the Speech Accessibility Project Challenge 2 (SAPC2) with a public speech recognizer that we fine-tuned on the challenge data. Our best entry scored 7.81% character error rate on Test1, the first of two hidden test sets (official baseline 15.26%). Our live entries ran on CPUs with a combined median latency of about 0.9 seconds at 7.96-8.07% error. On new speakers, fine-tuning reduced errors most for the speakers that an off-the-shelf recognizer finds hardest to understand. Sentences already seen in training had far lower error rates than new ones (2.58% vs 6.42%), a larger difference than being a familiar speaker made. We close with lessons for evaluating such systems.
Long-term time-series forecasting supports decisions in energy, transportation, and other applications where reliable predictions are needed beyond the observed data. Although forecasting research has advanced model architectures, fixed activation configurations provide limited guidance on how activation choices should vary across datasets and prediction horizons. This study aims to establish when task-specific activation choices improve forecasting accuracy and how previous experiments can guide the selection of suitable configurations for new tasks. We develop a calibrated selection policy that learns a reliable default from prior experiments, adapts the activation function and its location to each task, and retains the default when evidence for a change is insufficient. In an evaluation supported by 1,225 Informer training runs across 17 dataset variants and 51 forecasting tasks, the policy reduces mean squared error by 7.86% on average relative to the conventional fixed-activation baseline and improves 72.55% of tasks. These findings establish activation choice and placement as practical opportunities for improving existing forecasting models and show how accumulated experimental evidence can guide configuration decisions without repeating an exhaustive search for every task.
Hao Xue, Peng-Fei Zhang, Ying Cheng, Lanyun Zhu, Lei Zhu, Heng Tao Shen
Egocentric observations provide a robot-centric view of manipulation processes, capturing contact interactions and object-state changes that are important for visuomotor policy learning but difficult to observe from external cameras. However, collecting such data at scale remains expensive due to additional hardware requirements and platform-specific deployment costs. Synthesizing egocentric observations from large-scale exocentric manipulation videos provides a scalable alternative. However, existing methods struggle to synthesize physically consistent and policy-relevant egocentric videos under substantial viewpoint transformations, severe occlusions, complex object interactions, and inherently sparse observations from available exocentric viewpoints. Code will be made publicly available at https://github.com/xueha0/AuroraWorld
Yuan-Chiao Cheng, Yi-Hsuan Yang, Vincent K. M. Cheung
Music production typically involves multiple editing operations that each require specialized tools and expertise. Existing text-guided music editing systems support few operations or instruments, operate on short clips, or handle only one operation per instruction, unlike real workflows where several changes are requested at once. Here, we present ProCraft, a lightweight 0.5B-parameter autoregressive codec language model that edits 30s music excerpts end-to-end from free-form instructions, including compound requests. ProCraft supports seven operations across eight instruments, from tasks with a unique target (e.g., track removal) to creative ones (e.g., accompaniment generation). We align intermediate representations with pretrained music embeddings and pitch and instrument targets to preserve content outside the edit, and gated cross-attention in every decoder layer to keep generation tied to the instruction. To train on compound requests, we introduce On-the-fly Editing-pair Synthesis, which generates paired instructions and audio combining track-level operations with effect changes. ProCraft performs competitively against task-specific systems on single-task editing, and outperforms four free-text editors on compound instructions for track removal, extraction and denoising on real (MoisesDB) and synthetic (Slakh) recordings. In a user study (16 listeners), ProCraft was rated highest on removal and denoising and on-par with a 16-times larger model on extraction and instrument transfer. Ablations further show that representation alignment and cross-attention reduce leakage from unwanted tracks and raise editing success rates, especially on out-of-domain data.
Hengtong Shen, Li Yan, Hong Xie, Yaxuan Wei, Xinhao Li, Wenfei Shen, Peixian Lv, Fei Tan
Remote sensing (RS) change detection is essential for interpreting surface dynamics. Semantic change detection (SCD) further enables pixel-level understanding of multi-class transitions, yet remains sensitive to pseudo-changes induced by imaging conditions. Recent RS foundation models extract semantically consistent features across temporal and environmental variations, which is critical for mitigating pseudo-changes. However, existing SCD methods are often rigid and backbone-specific, lacking the flexibility to integrate diverse multi-scale features from emerging foundation models. To this end, we introduce a modular Cascaded Gated Decoder (CG-Decoder) that bridges various backbones and SCD tasks, processing multi-scale features in a coarse-to-fine manner while enabling adaptive change extraction. Building upon the RS foundation model PerA, we present PerASCD, a unified SCD framework. We further propose a Soft Semantic Consistency Loss (SSCLoss) to mitigate numerical instability in mixed-precision training. Extensive experiments on SECOND and LandsatSCD show that PerASCD achieves new state-of-the-art Sek scores (26.11% and 65.21%), surpassing the previous best by 0.61% and 4.95%, respectively. It also demonstrates exceptional data efficiency (outperforming the full-data baseline with 50% data), seamless cross-backbone generalization, and enhanced interpretability. Our approach maintains robust semantic consistency under radiometric variations, providing a reliable SCD solution. Code: https://github.com/SathShen/PerASCD.git.
Qiming Shi, Yikun Li, Beidi Luan, Rui Sun, Sinuo Wang, Delin Wang, Yan Ma, Jing Li 等 10 位作者
Large language model (LLM) agents are extending agent-mediated commerce on the Web from shopping for consumers to operating businesses for sellers. In online-to-offline (O2O) commerce, decisions are issued through Web-based platforms but realized through costly physical execution. Evaluating such agents requires going beyond individual transactions to ask whether decisions made online turn into sustained economic returns. We introduce AutoVend, a benchmark in which one LLM agent manages 20 vending machines across five site types through a Web-based operating interface for 180 simulated days. Its data-grounded simulator combines a shared warehouse, procurement lead times, paid replenishment, inventory expiration, and recurring operating costs. We evaluate five LLM agents against a rule-based inventory-control baseline and diagnose their site coordination, execution, and performance over time. Four of five evaluated LLM agents fall short of this baseline, and returns per dollar of inference cost rank models differently from returns alone. Operational analyses show that gains in individual metrics do not consistently yield stronger economic growth, as inventory and execution costs erode these gains. Trajectory analysis further reveals difficulties in grounding feasible actions, managing inventory commitments, and converting feedback into effective policy revisions.
We present \textbf{WeLM-Image}, a unified image foundation model that integrates multimodal understanding into text-to-image generation, instruction-guided editing, and multi-reference synthesis. Designed for single-GPU deployment in WeChat's large-scale consumer applications, the model processes text, reference images, and target latents as one interleaved sequence. A Mixture-of-Transformers (MoT) backbone combines shared attention with independent understanding and generation parameters to limit optimization interference. Equipping each path with Mixture-of-Experts (MoE) modules provides sparse capacity, with 31.7B stored parameters and approximately 5.4B generation parameters activated per target latent token. A generation-oriented data engine combines a knowledge-grounded visual vocabulary with large-scale visual clustering to balance semantic coverage and editing tasks, while training-time evaluation guides iterative data refinement. Bilingual annotations and verified editing supervision support progressive training up to native 1.5K resolution. We optimize generation and editing separately with rubric- and checklist-based rewards, then consolidate the specialized policies through noise-aware on-policy distillation. In the reported overall comparisons, the 1536-resolution model ranks first among the evaluated open-source models on PRISM-Bench, Qwen-Image-Bench, and UnicBench. MeanFlow-based few-step distillation yields WeLM-Image-Turbo, which generates a $1024\times1024$ image in 4.98\,s on a single NVIDIA H20 GPU using eight guidance-free steps.
Xianjing Zhu, Wenqi Lou, Yuanbo Wen, Teng Wang, Lei Gong, Chao Wang, Xuehai Zhou
Large language models (LLMs) offer a promising path for directly generating functionally correct and synthesizable HLS-C from natural-language specifications. However, existing approaches often struggle to turn multi-round tool feedback and reusable HLS knowledge into effective revisions, limiting end-to-end generation success. We propose MementoHLS, an agentic HLS generation framework built around executable and falsifiable dual-memory-guided repair. MementoHLS combines an Executable Repair Contract Library (ERCL), which converts cross-task HLS knowledge into evidence-gated repair actions, with a Reviewer-Grounded Falsification Ledger (RFL), which converts task-local repair history into trajectory evidence to suppress ineffective routes and preserve validated progress. To ensure reliable outcome attribution, a Same-Candidate Reviewer binds Parse, Compile, Testbench (TB), and Synthesis (Synth) outcomes to the corresponding candidate, while a Single-Commit Arbiter commits at most one candidate per round. Evaluated on 94 HLS-Eval and 170 Bench4HLS tasks, MementoHLS improves same-candidate TB$\wedge$Synth pass@1/pass@5 from 41.7%/57.4% to 58.9%/68.1% on HLS-Eval, and from 15.5%/32.9% to 32.4%/50.6% on Bench4HLS, demonstrating substantial improvements in end-to-end HLS generation reliability.
Dingxian Wang, Zhe Fu, Jiaqi Deng, Chao Yin, Ganesh Prabhu, Karthik Gomadam Rajagopal, Darwin Yi, Huan Huo 等 12 位作者
Coding agents increasingly collaborate on long-term projects. When taking over a project, they need to recover currently applicable requirements from the interaction history and update them in response to new requests. However, most coding benchmarks do not include long interaction histories, and some evaluations that consider history still do not cover the full process from requirement reconstruction to project delivery. We therefore provide agents with the interaction history, current request, and pre-request project environment, ask them to make the requested changes, and assess takeover performance through delivery tests. Across six model-harness configurations and 37 targets eligible for executable validation, the mean complete-delivery rate under Full History is only 4.95%. We then use requirement atoms to independently diagnose a larger set of takeover targets, examining relevant information identification, pre-request state reconstruction, and current request processing. We further supply correct pre-request or target states to examine the remaining difficulties in requirement processing and project modification. The diagnostics show that none of the six configurations produces fully correct post-request requirement states, and a case illustrates omission of previously recovered constraints during updating. With correct target states supplied, the mean complete-delivery rate is 8.11%, with build and regression pass rates of 96.40% and 90.99%, but target-test success remains 9.01%. These results show difficulties in both forming target requirement states and implementing supplied targets in existing projects. Based on these findings, we introduce ReqMemBench, which combines requirement diagnostics with delivery validation to support systematic evaluation of project takeover.
We address two open questions in streaming PCA via Oja's algorithm: sharp operator-norm convergence for general rank under sub-Gaussian data, and distributional inference for the resulting subspace estimator. Existing convergence analyses, even in the rank-one case, either assume bounded data or leave non-vanishing remainder terms that prevent adaptation to a polynomially vanishing tail spectrum, while existing distributional results are confined to the rank-one case. Our convergence theory removes these remainder terms and yields a sharp rate. In the dense-tail spiked covariance regime, this rate matches the minimax rate up to logarithmic factors. More generally, we prove a matching lower bound, up to logarithmic factors, across both dense-tail and sparse-tail regimes under a mild nondegeneracy condition. The analysis yields a linearization of Oja's iterates, which in turn enables a high-dimensional Gaussian approximation for the general-rank subspace estimation error with an explicit limiting covariance. We also establish a row-wise Gaussian approximation over convex sets for the aligned difference, recovering prior rank-one results as special cases. For practical inference, we develop an online multiplier bootstrap algorithm and prove its consistency. Beyond streaming PCA, our techniques contribute to Gaussian approximation and bootstrap inference for nonconvex stochastic approximation.
AI coding agents are increasingly embedded in real-world software development, collaborating with human developers while gaining broader access to codebases and tools. This creates a new attack surface: an agent can exploit human trust to sabotage development, for instance by inserting malicious code to accomplish a hidden side task. Most prior work studies AI sabotage in AI-only settings, paying limited attention to the role of human oversight in detecting and mitigating such malicious behavior. To address this gap, we conduct the first large-scale study of human oversight in AI coding sabotage. Over 100 participants collaborate with one of four frontier models (Claude-Opus-4.6, GPT-5.4, Gemini-3.1-Pro, and MiniMax-M2.7) on a long-horizon coding task lasting around five hours, designed to mimic real-world workflows. We find that 83/88 (94%) of developers in the no-monitor conditions fail to detect sabotage, and our analysis of participant feedback attributes this vulnerability to minimal code review, plausible cover story, and overtrust in agents. We further test the effectiveness of a safety monitor in one condition: while the monitor reduces sabotage success, sabotage still succeeds in 9/16 (56%) of sessions with a correct monitor alert. Drawing on participant feedback, we offer actionable suggestions for better monitor design. This work complements existing AI safety research and highlights an urgent need for human-centric safety mechanisms that account for human factors, particularly in long-horizon, real-world development settings.
AI 速览：提出统一基准 GraphIP-Bench，在同一黑盒协议下系统评测图神经网络模型窃取攻击与所有权防御，并揭示防御在提取后代理模型上的失效缺口。
Graph neural networks (GNNs) deployed as cloud services can be *stolen* through *model-extraction attacks*, which train a surrogate from query responses to reproduce the target's behaviour, and a growing line of ownership defenses tries to prevent or trace such theft. The title of this paper asks two questions: *how hard is it to steal a GNN?*, and *can we stop it?* Prior work cannot answer either, because experiments use inconsistent datasets, threat models, and metrics. We introduce *GraphIP-Bench*, a unified benchmark which evaluates both sides under a single black-box protocol. It integrates twelve extraction attacks, twelve defenses spanning watermarking, output-perturbation, and query-pattern-detection families, ten public graphs covering homophilic, heterophilic, and large-scale regimes, three GNN backbones, and three graph-learning tasks, and it reports fidelity, task utility, ownership verification, and computational cost on shared splits, queries, and budgets. We further add a joint attack-and-defense track which runs every attack on every defended target and measures watermark verification on the resulting surrogate, which exposes the protection that a defense retains after extraction. The empirical picture is short: stealing a GNN is easy at medium query budgets and most defenses do not change this; several watermarks verify reliably on the protected model but lose most of their verification signal on the extracted surrogate, which exposes a gap that single-model evaluations miss; and heterophilic graphs are systematically harder to steal, while a cross-architecture mismatch between target and surrogate reduces but does not prevent extraction. We release *GraphIP-Bench* with reproducible scripts and configurations, and integrate the attacks and defenses into the PyGIP library.
Small language models offer a compact basis for shell simulation without executing submitted commands, but full-output decoding requires them to predict recurring utility structure alongside content that varies across executions. We introduce template gap reconstruction (TGR), which supplies reusable output literals interspersed with sentinels to be interleaved with autoregressively generated gaps. A template selector chooses which template to be reconstructed, with the empty template retaining full-output decoding. We construct a template bank from 18.7 million outputs collected across heterogeneous execution environments and train a template selector and five reconstruction models initialized from various small language model checkpoints on a target environment excluded from bank construction. Evaluation demonstrates that TGR improves output edit similarity (O-ES) over full-output decoding (FOD) when the optimal template covers more than 25\% of the template, while performing similarly for lower-coverage outputs. Additionally, TGR outperforms FOD for shorter gap sequences, with O-ES values of 0.964 and 0.812, respectively, while maintaining similar performance for longer sequences. TGR reduces the number of tokens generated by 26.3--38.3\% across used tokenizer families. These results support using a reusable output structure as a practical design choice for small shell simulators, improving response fidelity and shortening generation targets while keeping template selection and residual prediction central to response quality.
Machine-learning surrogates of linear peeling-ballooning stability now run in milliseconds, but they are mostly deployed as point estimators: they do not say how wrong they may be, or when a query lies outside what they were trained on. For a fusion company designing a new device, every equilibrium of interest is outside the training data of existing surrogates, so this is the case that matters. We describe a model-agnostic trust layer for stability surrogates, made of conformal error bars, out-of-distribution alarms that route doubtful cases to the physics code, and recalibration from a few simulations of a new design, and we test it with pre-registered hypotheses. On a large public dataset of JET-like equilibria with ideal-MHD growth rates, the error bars hold their nominal coverage in distribution and fall short under a shift in plasma shape; a few dozen simulations of the new shape, or routing the cases the alarms flag, restore it. On resistive BOUT++ surrogates, the bars stay calibrated on held-out runs, the alarm flags tokamaks absent from training, and a scenario scan is screened in seconds instead of hours. Across equilibria from the SPARC library and the DIII-D coil set, a surrogate trained on our own graded BOUT++ labels beats the baseline on unseen equilibria, while a surrogate trained elsewhere ranks the same cases backwards and its alarm declines them. Building our own neural-operator surrogate, the harness exposed two input defects whose fix closed its gap to a simple network. For every surrogate tested, the trust layer keeps coverage near target, with a bar width set by that surrogate's own error. Not every pre-registered hypothesis held, and we report those results and the limits of all of them.
Shreya K. Chari, John S. Vardakas, Kostas Ramantas, Christos Verikoukis
Resource management for distributed edge networks hosting different services has been handled efficiently using Reinforcement learning (RL). Reward generalization in complex workflows is still an open issue within the RL realm. Therefore, this work proposes RAGAM-AI, a hybrid framework that combines the capabilities of RL algorithms and Generative-AI (Gen-AI) for powering an Agentic AI workflow that optimizes resource allocation at the Multi-access Edge Computing (MEC) units. This framework further resolves the reward generalization issue by utilizing Large Language Model (LLM) for feedback. A Multi- Agent Soft-Actor Critic (MASAC) using MathΣtral as a reward oracle is implemented for managing resource allocations at the MEC. RAGAM-AI interprets system state information and evaluates multi-objective trade-offs guiding the resource orchestration loop over distributed edge server agents. The complexity of the framework is further brought down by incorporating surrogate reward model. The overall performance is evaluated against the single agent SAC, REINFORCE and random policies for benchmarking. On using RAGAM-AI, lower allocation deviations for CPU, memory and bandwidth resources are reported while minimizing the latency and data rate costs. A scalable AI as a feedback (AIAF) solution optimizing network resources with significant performance improvement is thus presented.
Olivier Tieleman, Samuel J. Bell, Ziyi Zhu, Ting Su, Elizabeth Kroll, Thomas D. Hull, Caitlin A. Stamatis
Depression is a leading cause of disability worldwide, yet validated questionnaires such as the Patient Health Questionnaire-9 (PHQ-9) often go uncompleted, yielding sparse, biased data. Users of AI mental health applications often describe their mood, sleep and functioning unprompted, producing a continuous record of symptoms. Here, we predict PHQ-9 scores from these conversations alone. As in many clinical language tasks, labels are scarce and imbalanced. We address both by fine-tuning Qwen 3.5 27B with a regression head and augmenting 3,111 ground-truth labels with pseudo-labels from a frontier reasoning model and iteratively trained intermediate models, expanding to 5,524 users with more balanced scores. On 422 held-out users, the model achieved MAE $= 3.98$, Pearson $r = 0.67$, and AUC $= 0.88$ at the clinical threshold (PHQ-9 $\geq 10$); AUC was $> 0.80$ across severity thresholds from PHQ-9 $\geq 3$ to PHQ-9 $\geq 24$, demonstrating that the model captures depression severity across the full clinical spectrum. These results support passive depression monitoring without self-report.