Machine Learning
HiRoute: Hierarchical Routed Prompt Tuning for Safety Alignment of Large Language Models
Fangzhou Chen, Shiji Zhao, Mengyang Wang et al. (+4 more)
Large
language models (LLMs) remain vulnerable to harmful requests and jailbreak attacks. Parameter-efficient safety
alignment methods based on
prompt tuning typically rely on a single global
prompt or externally selected
prompt modules. Such static designs struggle to maintain a cross-category safety boundary while generating constructive responses tailored to specific risks and avoiding over-refusal of benign inputs. To address these limitations, we propose HiRoute, an input-adaptive hierarchical
prompt-tuning framework that separates category-agnostic safety control from category-specific response guidance. HiRoute first trains a lightweight hierarchical router on representations extracted from a frozen LLM to jointly detect harmful intent and predict multi-label risk scores. It then freezes both the backbone model and the router and uses preference optimization with alternating gradient updates to learn a shared coarse-grained
prompt and a set of fine-grained
prompt experts as continuous
embeddings. At inference time, benign inputs bypass the safety branch, whereas risky inputs are processed using the shared
prompt together with a router-weighted mixture of risk-specific
prompt experts. Experiments across three instruction-tuned models show that HiRoute achieves high safety rates across multiple safety
benchmarks while preserving safe-response helpfulness, reducing over-refusal, and maintaining competitive performance on general-purpose tasks.
Read the full abstract ▾
All authors Fangzhou Chen, Shiji Zhao, Mengyang Wang, Qihui Zhu, Ranjie Duan, Maoxun Yuan, Xingxing Wei
Information Retrieval
Query Translation vs. Cross-Lingual Embeddings for Sinhala-Tamil E-Government Information Retrieval
Dharshi Balasubramaniyam and Tiroshan Madushanka
This paper presents a comparative evaluation of cross-lingual information
retrieval (CLIR) methods for retrieving English government information using Sinhala and Tamil queries. Two CLIR paradigms are investigated: Query Translation (QT), employing Google Translate, NLLB, and mBART50, and Cross-Lingual
Embeddings (CLE), using LaBSE,
multilingual E5, and BGE-M3, with monolingual English
retrieval as the baseline. Experiments are conducted on a human-verified
benchmark comprising 500 Sinhala, Tamil, and English question-answer pairs derived from 1,699 segmented contexts from Sri Lanka's Government Information Center (GIC).
Retrieval performance is evaluated using Recall@k (k = 1, 3, 5, 10, 15). Monolingual
retrieval performs poorly (Recall@15 <10%), whereas all CLIR approaches substantially improve
retrieval accuracy. Among them, BGE-M3 achieves the highest Recall@15, reaching 96.2% for Sinhala-English and 95.6% for Tamil-English, outperforming the best QT approach (Google Translate: 92.4% and 93.0%) while avoiding translation overhead. These results demonstrate that
multilingual embedding models provide a more effective and scalable solution for cross-lingual
retrieval-augmented generation (RAG) in low-resource government domains.
Read the full abstract ▾
Computers and Society · Cross-listed from math.HO
Fundamental Mathematics in the Age of AI – The Residue, the Journey, and the Ecology
Benjamin Collas
Large
language models have begun refuting long-standing
conjectures and solving long-open problems. The introspection this has prompted about the future of mathematical discovery is well under way, and the anxiety accompanying it legitimate – but both, we claim, are attached to the wrong loss. What machines now produce is the countable part of mathematics –
theorems, proofs, refutations – which was always the work's residue, not its product. The distinction is old, and not economic: a result can be taken in its finished essence, or in the operations that engendered it. The product is human understanding: not a stock of results but a collective, hard-won way of deciphering the world and acting upon it. The two are arcs of a single loop: understanding tells us where to look; looking produces the residue; and taking it up again, one journey at a time, rebuilds shared understanding. Machines are strong on the countable arc, absent from the one that feeds it. The peril is to leave the loop open. AI did not create the confusion between residue and product; it has called a bluff long on the books, driving the cost of the residue towards zero and making the scarce thing visible at last. A new instrument makes a new way of working before it makes a new result. The pressing questions are therefore institutional: who can check an announced result, whoever announces it; what work and training become for the next generation of researchers; and whether the one thing that cannot be mass-produced – the journey that nourishes a shared understanding – continues to be funded. Mathematics, we argue, is uniquely placed among the sciences on the first – a proof answers to no one's permission – and uniquely exposed on the other two: teaching cannot go on as before, and no collective position yet exists; and the journey has never had a price our institutions knew how to pay. The decision is ours.
Read the full abstract ▾
Computation and Language
FastThaiG2P: Lightning-fast Thai Grapheme-to-phoneme Conversion for Voice Agent Pipelines
Charin Polpanumas
FastThaiG2P provides sub-millisecond Thai grapheme-to-phoneme conversion for text-to-
speech pipelines (International Phonetic Alphabet and Kokoro-TTS conventions) using a PyThaiNLP-tokenized, extensible dictionary and normalization rules for common Central Thai
speech. The approach achieves an average latency of 0.15 ms per utterance on a
benchmark of 27,242 synthetically generated utterances, of which 30% is spent on tokenization, 12% on normalization, and 58% on out-of-vocabulary fallbacks (0.5% OOV rate). To demonstrate its effectiveness, we used FastThaiG2P to phonemize Som-TTS, an open
dataset containing 20 hours of grapheme-and-audio pairs, then trained an 82M-parameter StyleTTS 2 model based on a Kokoro-TTS recipe. The resulting model vocalizes intelligible Thai
speech suitable for prototyping and development at 0.25 real-time factor (4x real-time) with ONNX inference on CPU.
Read the full abstract ▾
Information Retrieval
A Comprehensive Empirical Evaluation of Vector Database Systems for Approximate Nearest Neighbor Search: Performance, Quality, and Resource Trade-offs
Ashen Rashmiks and Tiroshan Madushanka
Vector databases have emerged as critical infrastructure for modern artificial intelligence applications, particularly
retrieval-augmented generation (RAG), semantic search, and recommendation systems. Despite their growing importance, there remains a significant gap in comprehensive, reproducible
benchmarks that jointly evaluate
retrieval quality, query latency, throughput, and resource utilization. We present a systematic empirical evaluation of seven prominent vector database systems: FAISS, Qdrant, Milvus, Weaviate, Chroma, pgvector, and LanceDB. Our methodology spans six diverse
datasets, from classical computer-
vision descriptors (SIFT, GIST) to
transformer-based text
embeddings (MS MARCO, GloVe), encompassing over 4 million vectors at dimensionalities from 96 to 960. We measure 15 metrics spanning
retrieval quality (Recall@K, Precision@K, MRR, NDCG@K, Hit Rate@K), query performance (latency percentiles, QPS, cold-start latency), and resource consumption (index build time,
memory, storage). On SIFT1M, FAISS achieves the highest single-node throughput (866 QPS) but lacks database operational features; Weaviate provides the best out-of-the-box recall (> 99%); Qdrant offers the best latency among full databases (4.55~ms median); and LanceDB trades
retrieval quality for substantially faster index construction. We derive system-selection guidelines for practitioners and release our benchmarking framework as open-source software.
Read the full abstract ▾
Computer Vision
Structured Local Differential Modeling for AI-Generated Image Detection
Jiazhen Yang, Ruijin Jin, Junjun Zheng et al. (+3 more)
The rapid advancement of AI-generated content has made the reliable detection of generated images an increasingly critical challenge. Existing detection methods are often dominated during training by semantically salient components with high signal-to-noise ratios (SNRs), thereby suppressing subtler forensic cues associated with the underlying generation mechanisms and embedded in low-level statistical structures. From an information-theoretic perspective, we present a key insight: effective detection in the low-level statistical space requires mitigating the dominance of semantic components while emphasizing and amplifying responses to low-SNR forgery traces. Building on this insight, we propose RippleNet, an AI-generated image detection framework based on local differential signals. RippleNet adaptively identifies forgery-sensitive regions and constructs multi-directional, multi-scale differential representations within local neighborhoods, explicitly characterizing anomalous patterns in neighborhood statistics. More importantly, we refine the
attention mechanism to operate within the local differential representation space, enabling the model to establish explicit dependencies at a finer statistical granularity. This design facilitates the capture of pixel-level forgery traces that are difficult to model using conventional convolutions or image-wide patch-level
attention. Extensive experiments on multiple public
benchmarks and under cross-generator evaluation settings demonstrate that RippleNet achieves consistently competitive performance.
Read the full abstract ▾
All authors Jiazhen Yang, Ruijin Jin, Junjun Zheng, Xiangheng Kong, Zunlei Feng, Jie Lei
Computer Vision
Erase but Preserve: Controllable Removal of Copyrighted Animation Characters via Optimized Semantic Anchors
Qiao Li, Xiaomeng Fu, Wangjia Yu et al. (+4 more)
The exceptional generation capabilities of text-to-image
diffusion models have raised copyright concerns, particularly the unauthorized reproduction of animation characters. Existing concept erasure methods fall short for animation character erasure: model modification methods struggle to identify suitable anchors for diverse, highly distinctive characters;
prompt-based steering methods lack fine-grained control for precise intervention. These approaches often yield incomplete erasure and degraded image fidelity, hindering real-world deployment. In this paper, we propose a controllable method operating on the model's continuous textual representation to erase target characters during generation. We optimizes an anchor
embedding via structural and detailed constraints to serve as a character surrogate, then replaces target-related
embeddings with the anchor via a structure-aware adaptive strategy. Experiments show that our method achieves state-of-the-art erasure effectiveness and image fidelity preservation, while supporting controllable erasure degree, multi-target removal, and model transferability. Moreover, our optimized anchors are plug-and-play with current model modification baselines to improve their erasure performance.
Read the full abstract ▾
All authors Qiao Li, Xiaomeng Fu, Wangjia Yu, Runze He, Baisen Wang, Jiao Dai, Jizhong Han
Machine Learning
CoMedBench: A Multi-Source Benchmark of Synthetic Medical Data Fidelity and Downstream Utility
Akanta Das, Farhad Al-Amin Dipto, Mrinmoy Sarkar Anto et al. (+3 more)
Access to clinical data is essential for developing reliable healthcare machine learning systems, but direct use of electronic health records is constrained by
privacy regulation, institutional review, data-use agreements, and the risk of re-identification. Synthetic data promises a practical alternative: it can preserve useful statistical and clinical structure while reducing exposure of sensitive patient records. Prior studies often evaluate a single generator, one
dataset, or a narrow downstream task, making it difficult to know when synthetic data can support model development and when it fails to preserve task-critical signal. We introduce CoMedBench, a reproducible
benchmark that evaluates a family of generators under a common clinical-validity framework and one shared training and evaluation engine, spanning static tabular and temporal downstream tasks on established critical-care
datasets. In total the
benchmark spans 37
dataset-task pairs across two modalities consists of 20 static tabular and 17 temporal ICU time-series-drawn from seven public data sources: three intensive-care databases (MIMIC-III, MIMIC-IV, and eICU) together with the UCI Machine Learning Repository, the CDC BRFSS diabetes cohort (2015), NHANES (1999-2014), and the pycox survival
datasets (GBSG and METABRIC). The
benchmark evaluates both statistical fidelity and task utility by comparing models trained and tested across real and synthetic data. In these settings, synthetic training data preserves most of the downstream signal: on tabular tasks the reference generator CoMed-CTGAN retains a mean AUROC utility (the synthetic-to-real performance ratio) of 90.6%, rising to 97.3% for the strongest generator, CoMed-TVAE. Temporal ICU tasks are harder and more generator-sensitive: CoMed-CTGAN retains 81.6% (AUROC) and only 64.0% under the imbalance-sensitive AUPRC, whereas CoMed-TVAE still retains ~95% (AUROC).
Read the full abstract ▾
All authors Akanta Das, Farhad Al-Amin Dipto, Mrinmoy Sarkar Anto, David Rehkopf, Ayin Vala, Tanmoy Sarkar Pias
Computational Engineering
Julia for CFD: A Critical Survey of Ecosystem, Performance, and Composability
Tianbai Xiao
Modern CFD increasingly places simulation inside workflows for design, inference, optimization, and data-driven modeling, creating pressure to connect physical models, numerical kernels, heterogeneous hardware, differentiation, and learning. Julia offers a distinctive approach: high-level scientific abstractions can be specialized for performance and composed within a common
language and compiler ecosystem. This critical survey examines where that model benefits CFD software and where its limits remain. We review representative open-source projects and synthesize application-level evidence on performance, scalability, accelerator portability, automatic differentiation, and software composition. Published results demonstrate credible Julia-native CFD on large distributed CPU systems and multi-GPU platforms, as well as emerging differentiable workflows. Comparisons with C++ performance-portability frameworks, finite-element domain-specific
languages, and JAX-based differentiable CFD show that these capabilities are not unique to Julia. Julia's distinction is their integration through shared types, dispatch, and specialization. The evidence is mixed: Julia has progressed beyond proof of concept in several CFD regimes, but still lacks the ecosystem breadth, industrial tooling, and deployment experience of established C/C++/Fortran environments. Its strongest current role is as a platform for developing and testing CFD architectures that connect simulation with downstream analysis.
Read the full abstract ▾
Machine Learning · Cross-listed from physics.ins-det
Fine-tuned Normalizing Flows for ALICE Zero Degree Calorimeter Fast Simulation
Emilia Majerz, Jacek Otwinowski, Witold Dzwinel et al. (+1 more)
Simulating the ALICE Zero Degree Calorimeter (ZDC)
neutron detector responses at the LHC is computationally expensive, requiring complex Monte Carlo chains. We develop a generative surrogate, focusing on Normalizing Flows (NFs). Through transfer learning, we pre-train on the full imbalanced
dataset and fine-tune specialized models for different particle types ($\gamma$, $n$, $\Lambda$, $K_S^0$, $\Sigma^+$) using two gradual-unfreezing schemes. As standard ZDC metrics like Wasserstein distance overlook conditional structure, we introduce refined metrics: conditional weighted MAE, dispersion ratio, and Jaccard co-activation error, that better capture physics-relevant input-output dependencies and response variability. Our ensemble of fine-tuned models achieves a Wasserstein distance of $1.61 \pm 0.02$, outperforming baselines across all metrics. This work provides a generalizable NF-based framework for LHC detector simulation, combining NFs, conditional fine-tuning, and physics-motivated evaluation.
Read the full abstract ▾
All authors Emilia Majerz, Jacek Otwinowski, Witold Dzwinel, Jacek Kitowski
Human-Computer Interaction
Considering Contribution Statements in Visualization and HCI Research
Mara Solen, Wesley Willett and Andrew M McNutt
Contribution statements are an increasingly common way to make research labor visible, reduce academic malfeasance, and provide broader transparency. Despite this potential value, they remain uncommon in visualization and HCI. To explore this gap, we conducted an online study with (N=21) visualization and HCI researchers. We find a range of differing opinions about the utility of contribution statements, which are set against a background of tensions relating to contribution frameworks that inadequately fit contribution types in HCI and especially visualization, power dynamics between authors, bias in authorship perceptions, and the tedium of providing yet another form of documentation. From these factors, we offer a modest recommendation to authors: consider contribution statements. There are contexts when they may usefully explicate work, and others where they can cause author-team conflict or become a burdensome chore. Regardless of whether they are used by readers or not, we suggest that scaffolded mechanisms for reflecting on contribution roles are valuable both for public accountability and internal
alignment. To institutions, we recommend that contribution statements be exempt from page or word limits, and that flexible templates and examples be provided to authors, but that they continue to not be required. By surfacing these perspectives, we seek to open a dialogue about what constitutes authorship, how our community might move toward more equitable and transparent attribution practices, and where the visualization and HCI communities might be uniquely equipped to help.
Read the full abstract ▾
Information Theory · Cross-listed from cond-mat.stat-mech
Thermodynamics of Learning: A Typed Four-Component Accounting of Memory, Fit, and Value
Akihito Sudo
What a finite learning device has recorded and what will hold value for it on future tasks are not the same quantity. We develop a typed accounting for finite-state learning devices that separates four components: a training-side fit functional $\Phi_{\mathrm{fit}}$, the record-correlation stock $J_{D}=I(M;D)$, an update-side search ledger $\sigma_{M}$, and an operational capital value $V(M;T,b)$. This value is the work gap between an informed protocol class and a blind class obtained by deleting the
memory-read port and re-optimizing from scratch. (I) Separation: for every $n$, there is a device family on which record correlation and world correlation grow by $n\ln 2$ while the capital gain is exactly zero. In the $\mathrm{flat}^{*}$ regime, data-free updates never increase $V$. (II) Capitalization ledger: an exact $\mathrm{flat}^{*}$ extraction identity and a universal ledger identity give, for (F5$'$)-stable $M$-local updates under a no-discarded-record-correlation condition (f), the bound $\eta_{\mathrm{cap}}\le 1$ for the capitalization efficiency $\eta_{\mathrm{cap}}=\Delta V/(k T\,\sigma_{M})$, together with necessary and sufficient conditions for equality. (III) Value retention: for the retention gap $L_{\mathrm{gen}}$ and retention ratio $\rho_{\mathrm{gen}}$ (the former carries no sign constraint; the latter is defined for positive training-side value and is not confined to $[0,1]$) we give a two-layer
alignment domain: an exact exchange rate between value and the side-information-adjusted record fit $I(M';D\mid Y)$ without any record-side-information independence assumption, and a raw record-stock exchange rate under a joint side-information neutrality condition $(M,D)\perp Y$, whose boundary is marked by an explicit one-time-pad witness. These are statements about finite-device value retention under task-distribution shift, not a theory of statistical generalization.
Read the full abstract ▾
Cryptography and Security
PIPES: Securing Agent Perception with Provenance and Priors
Sanjay Kariyappa, Severin Klingler and G. Edward Suh
Tool-using
agents consume external data from sources with different levels of trust, yet tool responses rarely identify who produced each component or what it should convey. We show that this gap enables state-corruption attacks, in which attacker-controlled content makes environmental claims beyond the informational authority of its response component and corrupts the
agent's perceived environment, making the resulting action appear justified to existing guardrails. We introduce PIPES (Provenance-Informed, Prior-Enforced Screening), which screens response units using semantic priors and source provenance. PIPES uses static field contracts when schemas provide stable expectations, and conditions screening of open-ended content on the pre-response trajectory and trusted provenance metadata. It marks units that violate their semantic prior or the provenance hierarchy; deployments may remove, warn, block, or escalate detected violations. We instantiate atomic removal and evaluate PIPES against adaptive PAIR-style attacks. Across the three VitaBench and three AgentDyn splits with Gemma 4 31B IT as the target
agent, PIPES reduces average attack success from 84.7% to 2.3%, while preserving average benign utility (92.5% with PIPES versus 90.6% without defense).
Read the full abstract ▾
Artificial Intelligence
ARAC: Benchmarking Auto-Research's Alignment and Completeness on End-to-End Researchs
Jiale Cui, Yueyao Yuan, Kaixi Zhong et al. (+3 more)
The rapid advancement of Auto-Research has surfaced a fundamental evaluation challenge: how can we measure the
alignment, logical coherence, and evolutionary completeness of its research trajectory with human research behavior? We propose Auto-Research's
Alignment and Completeness, ARAC-Bench: a Researcher-Mimicking Evaluation framework that shifts the objective from matching final answers to reproducing high-quality human research processes. The framework operates through two synergistic components: the Academic Cognition Skills system, which is the first to transforms implicit reviewer expertise into stage-calibrated, quantifiable rubrics; and a three-stage capability diagnostic protocol, which decomposes the research process under strict modular constraints into three traceable, mutually independent dimensions: Proposal, Experiment, and Synthesis. Systematic evaluation of 11 SOTA frameworks yields a best
alignment score of only 67.9 of 100, revealing a significant gap in simulating rigorous human methodology. Validation against Ph.D. Candidates rankings shows a strong correlation of 0.8141, confirming that ARAC-Bench reliably reflects the dimensions researchers truly value. ARAC-Bench provides not only a fine-grained diagnostic tool but also a scalable reward signal for training the next generation of autonomous research systems.
Read the full abstract ▾
All authors Jiale Cui, Yueyao Yuan, Kaixi Zhong, Xiaogang Xu, Jiafei Wu, Zhe Liu
Computer Vision
Beyond Correctness: Benchmarking and Aligning Response Behaviors in Hybrid-Thinking MLLMs
Xinming Wang, Weinong Wang, Hongming Yang et al. (+13 more)
Hybrid-thinking
multimodal large
language models (MLLMs) allow a single model to alternate between deliberative thinking and latency-efficient non-thinking inference. Although these modes differ in
reasoning budget, their delivered responses should satisfy the same user-facing standard. Correctness alone may not characterize this response quality; we therefore evaluate task accuracy and response-pattern failures as complementary outcomes. We study this gap through response-pattern
alignment: whether thinking and non-thinking interfaces preserve acceptable final-response behavior. We introduce PatternEval, a failure-enriched diagnostic
benchmark comprising 2,415
multimodal prompts spanning visual perception and grounding, structured image understanding, and
multimodal knowledge
reasoning. PatternEval tests four recurrent failures: chain-of-thought leakage, response repetition, logical contradiction, and performative
reasoning. Response-pattern failures are widespread across models from different providers, with non-thinking inference exhibiting substantially higher failure rates and thereby creating systematic misalignment between thinking and non-thinking interfaces. Motivated by this diagnosis, we develop PatternRM, a response-level reward model, and PatternRL, which introduces pattern-specific penalties during reinforcement learning. Experiments on Qwen3-VL-4B and Qwen3-VL-8B show that incorporating pattern-specific penalties into reinforcement learning can mitigate cross-mode misalignment while incurring a marginal task performance trade-off. Together, PatternEval and PatternRL provide an evaluation-and-training framework for aligning user-visible response patterns across hybrid-thinking interfaces.
Read the full abstract ▾
All authors Xinming Wang, Weinong Wang, Hongming Yang, Yansong Lin, Zheng Ruan, Shangpin Peng, Qiming Peng, Nan Qiao, Fengyuan Lu, Guoqing Ma, Marito Li, Songyang Zhang, Saiyong Yang, Han Hu, Yonglong Tian, Xu-Yao Zhang
Computer Vision
SCOPE: Subspace Clustering with Online Per-Head Top-K Estimation for Sparse Video Attention
Qi Zhao, Qirui Li, Hanlin Tang et al. (+10 more)
Diffusion Transformers (DiTs) incur quadratic self-
attention cost over spatiotemporal
tokens. Existing training-free sparse
attention methods often construct sparse masks from block-level or cluster-level proxy scores, which can obscure fine-grained differences among keys and miss high contribution keys under aggressive sparsity. Moreover, such proxy scores may yield overly concentrated softmax distributions, causing Top-$p$ to retain too few keys for some query clusters. Although a fixed Top-$k$ minimum alleviates this failure mode, a shared value cannot adapt to variations across heads and inputs. To address both limitations, we propose SCOPE, a training-free sparse
attention framework that combines 3D-RoPE-aligned key subspace clustering with online per-head Top-$k$ estimation for efficient
video-DiT inference. SCOPE partitions post-RoPE keys into temporal, height, and width subspaces, clusters them independently, and aggregates the corresponding centroid scores through lookup tables to obtain per key proxy scores for each query cluster. Building on existing hybrid Top-$p$/fixed Top-$k$ selection, SCOPE derives a head-specific Top-$k$ value online by averaging the initial retained key counts within each head, weighted by query cluster size, and selects additional keys only for query clusters whose initial retained key counts fall below this value. Sparse
attention is then computed over the selected original keys and values. Across six model–task configurations, SCOPE consistently outperforms existing training-free baselines in both fidelity and latency, achieving up to a $1.99\times$ end-to-end speedup on 720p HunyuanVideo with $28.46$ dB PSNR relative to dense
attention.
Read the full abstract ▾
All authors Qi Zhao, Qirui Li, Hanlin Tang, Yiduo Li, Zhen Guo, Cuifeng Shen, Chao Xu, Zhaosheng Chi, Xiaojin Lu, Kan Liu, Tao Lan, Lin Qu, Xi Li
Computation and Language
CRAFT: LLM-Based Iterative Refinement for Temporal Reasoning over Clinical Narratives
Chengyang He, Tahreem Arif, Marko Zivkovic et al. (+3 more)
Understanding the temporal progression of symptoms in clinical narratives is critical for disease monitoring, safety surveillance, and causality assessment. Clinical narratives, however, rarely provide explicit temporal anchors. Current approaches to temporal information
reasoning focus predominantly on pairwise relation classification across multi-visit and timestamp-rich records, leaving the reconstruction of structured symptom trajectories from individual anchor-sparse reports largely unaddressed. We propose CRAFT, an LLM framework that pairs a generator with a constraint-based verifier to iteratively produce and refine stage-wise symptom timelines through targeted feedback. We conduct evaluation on MedTempo, a new
benchmark of 5,347 vaccine adverse-event narratives spanning three COVID-19 vaccine types, with expert-validated temporal stage annotations for 3,166 reports. Experiments across four LLM backbones demonstrate that CRAFT consistently improves temporal ordering accuracy, with ablation analysis isolating the contribution of generator and verifier components across model capability levels.
Read the full abstract ▾
All authors Chengyang He, Tahreem Arif, Marko Zivkovic, Lijing Wang, Yue Ning, Ping Wang
Information Retrieval
DrEM: Dual-Side Robust Ensemble Ranking from Noisy User Preference Predictions in Video Recommendation
Canwei Huang, Tiantian He, Xiaoxiao Xu et al. (+5 more)
Industrial
video recommendation systems typically adopt a multi-stage architecture. At the ensemble ranking stage, multi-dimensional user preference predictions (pxtrs) from an upstream multi-task model are fused into a unified ranking score to reflect user satisfaction. Since users' true satisfaction is difficult to observe directly, ensemble ranking models commonly use pxtrs both as input features and as a source for constructing proxy preferences. However, as outputs of an upstream prediction model, pxtrs inevitably contain prediction noise, which propagates to downstream learning across two sides. On the supervision side, noisy pxtrs may flip proxy preferences and introduce erroneous gradients. On the feature side, pxtr noise may propagate through model inputs and destabilize ranking scores. Existing ensemble ranking methods typically treat pxtrs as reliable signals and overlook such prediction noise. To address this, we propose DrEM, a dual-side robust ensemble ranking framework. Our DrEM introduces a risk-denoising robust loss that corrects the empirical risk using estimated preference flip probability. Meanwhile, it samples perturbations from the distribution of prediction noise and introduces a preference-preserving ranking consistency regularizer to improve feature-side output stability. Theoretically, we obtain an approximate distribution of the prediction noise and prove that the robust loss remains superior under flip probability estimation error. Extensive offline experiments and large-scale online A/B tests demonstrate the effectiveness and robustness of our DrEM.
Read the full abstract ▾
All authors Canwei Huang, Tiantian He, Xiaoxiao Xu, Jun Zhang, Ziran Deng, Weike Pan, Chunjie Chen, Kaiqiao Zhan
Computation and Language
ViTOED: A Dataset for Target-Oriented Emotion Detection on Vietnamese Social Media Texts
Chanh Vo, Son T. Luu and Ngan Luu-Thuy Nguyen
This paper introduces ViTOED, a novel
dataset for target-oriented
emotion detection in Vietnamese social media texts. The ViTOED comprises 10,985 user comments and 21,244 manually annotated opinion quadruples (source, target, expression, polarity) that follow strict guidelines. The
dataset reveals Vietnamese-specific phenomena, such as implicit sources and targets and vocabulary ambiguities, enabling deeper analysis of user
emotions toward entities. We propose a baseline using structured sentiment graphs and evaluate various Vietnamese pre-trained
language models. The empirical results highlight challenges in span detection and relation extraction and indicate substantial room for model improvement in Vietnamese Target-Oriented
Emotion Detection tasks.
Read the full abstract ▾
Computer Vision
CW-BASS v2: Saturation-Aware Pseudo-Label Selection for Semi-Supervised Segmentation under Foundation-Model Teachers
Ebenezer Tarubinga
Semi-supervised semantic
segmentation has long turned on one question, which pseudo-labels to trust, and a generation of selection rules, dynamic thresholds, per-class curricula, soft confidence weights, answered it for the noisy, under-confident ResNet teachers of their day. Self-supervised foundation encoders change the regime: with a DINOv2 teacher, confidence saturates, so the filtering that helped a weak teacher can hurt a strong one. We propose CW-BASS v2, a saturation-aware pseudo-label selection method that reads the teacher's confidence regime rather than committing to one rule. It pairs held-out calibration, an unbiased per-class noise estimate, with a self-adaptive confidence floor that provably bounds retention away from 1, and combines them in a one-pass gate: measure the reliability of the teacher's confident set, pi_kept = Pr[correct | c >= tau], on a held-out slice, and filter strictly when it meets the confidence demanded (pi_kept >= tau), falling back to the adaptive floor otherwise. The boundary is the pre-existing operating threshold, not a value tuned to mIoU, and across six DINOv2 teachers it makes the correct strict-vs-floor call blind. CW-BASS v2 thus recovers the UniMatch V2 operating point on the saturated
benchmarks by selecting strict (Pascal VOC 1/8 87.4 against its reported 87.9; Cityscapes within 0.5), and improves on it where the confident set is unreliable (pi_kept ~ 89%, ADE20K), where the floor edges ahead (+1.5 mIoU, single seed). The gate is principled because the failure it avoids is measured, not assumed: on a reliable, saturated teacher the confidence distribution's dynamic range collapses (98% of Pascal pixels >= 0.95), so an adaptive cutoff floods the retention mask and self-training decays into confirmation bias.
Read the full abstract ▾
Neural and Evolutionary Computing
Insights from Multi-tasking the EAX Algorithm for the Travelling Salesperson Problem
Liam Wigney, Aneta Neumann, Yew-Soon Ong et al. (+1 more)
Evolutionary multitasking allows several related problems to be solved in a single run of an
algorithm. In this paper, we investigate integrating evolutionary multitasking with Edge Assembly Crossover (MT-EAX) to solve the classical Travelling Salesperson Problem (TSP). To fairly compare MT-EAX against standard EAX under strict compute budgets, we evaluate three scaling methods: generation scaling, population scaling, and balanced scaling. Our results show that generationally scaled MT-EAX is highly effective compute-wise in the early stages of the search, saving $60\%$ to $90\%$ of compute for equal or better solution quality. We observe that instance geometry has a significant impact, with clustered, normally distributed instances securing larger improvements than uniformly distributed ones. However, when scaling by population or utilising explicit solution transfer, the results are negative due to population starvation and incompatible cross-instance parent selection. We demonstrate that the advantage of MT-EAX derives from increased diversity through parallel search in early generations, which can be successfully preserved using a decoupled configuration to often strictly outperform or match standard EAX performance at final convergence.
Read the full abstract ▾
All authors Liam Wigney, Aneta Neumann, Yew-Soon Ong, Frank Neumann
Software Engineering
Memorization Diagnostics for Code LLMs Should be Scale-Aware
Prateek Kumar Rajput, Abdoul Aziz Bonkoungou, Alberick Euraste Djiré et al. (+6 more)
The extent to which large
language models for code rely on memorization over genuine understanding remains highly debated. While current literature frequently reports widespread memorization, evaluating the underlying probing techniques across dense architectures reveals a severe breakdown in their utility at scale. Traditional encoder-style probes using perturbations such as synonym fuzzing or dead-code insertion struggle to expose memorization in scaled models, even on known-contaminated
benchmarks, and decoder-style probes that rely on log probabilities show similar performance degradation. The specific mode of failure for these probes, particularly why such techniques disrupt smaller models but fail to impact larger ones, motivates us to untangle representation load from memorization rather than treating them as a single phenomenon. By applying invertible mathematical transforms to numeric problems, we isolate these two factors and reveal that scaled encoders successfully absorb substantial representation load while still converging on the correct family of solutions. In practical software engineering, this ability to adapt to varying surface forms is what truly matters for usability and generalizability in LLM and agentic applications. Whether a specific solution was seen during training becomes a much less pressing question because although memorization inflates scores on contaminated
benchmarks, factoring out representation load makes it debatable how much we should truly care if a functional answer was originally memorized. Future evaluations must therefore be built around separating these phenomena rather than relying on methodologies that quietly entangle them.
Read the full abstract ▾
All authors Prateek Kumar Rajput, Abdoul Aziz Bonkoungou, Alberick Euraste Djiré, Xunzhu Tang, Yewei Song, Iyiola Emmanuel Olatunji, El Hacen Diallo, Jacques Klein, Tegawendé F. Bissyandé
Computers and Society
A Generative Framework for the Creation of Multi-Attribute Geographically-Explicit Synthetic Population
Jinlin Wu, Si Qiao, Yi Liu et al. (+2 more)
Generating multi-attribute synthetic populations with realistic joint distributions and geographic variation is a foundational requirement for geo-simulation techniques, such as micro-simulation and
agent-based modeling. However, it remains challenging for existing methods to reconstruct region-specific joint distributions from aggregated-level data alone. Thus, we propose a hierarchical
diffusion-based generative framework that utilizes a realistic region-specific joint distribution of multiple attributes as the training target to create a synthetic population along with assigning their explicit home and work locations. Applied to 50 U.S. states and Washington, D.C., this framework generates a nationwide geographically-explicit synthetic population consisting of 332,387,543 individuals with five attributes (e.g., age, gender, employment, education, income). Held-out regional experiments show improved reconstruction of joint distributions relative to Iterative Proportional Fitting (IPF) and a one-shot
diffusion baseline. At the same time, the location assignment preserves major residential and workplace patterns. As such, the proposed framework provides a scalable generative approach for creating geographically explicit synthetic populations at both regional and national levels. By reconstructing region-specific joint distributions of these five attributes using this framework, the resulting synthetic population could introduce more realistic behaviors into geo-simulations, such as
agent-based modeling, enabling further exploration of the emergence of complex urban phenomena through human interactions.
Read the full abstract ▾
All authors Jinlin Wu, Si Qiao, Yi Liu, Fuzhen Yin, Na Jiang
Computer Vision
PatchGen: Learning Soft Intra-Image Predictive Subsets for Visual Generalization
Zhaorui Tan, Weimiao Yu and Xi Yang
Visual classifiers are expected to generalize under data shifts, target shifts, and their combinations, yet most existing methods focus on domain invariance while failing to address intra-image predictive sufficiency. We investigate the structural hypothesis that each image contains a sample-adaptive oracle intra-image predictive subset sufficient for label prediction, while the remaining patches form non-essential complementary context that may correlate with the label. The theoretical analysis shows that restricting prediction to this oracle subset preserves the Bayes risk achievable by the full-patch representation while admitting a complexity bound that tightens with the oracle-subset size. Based on this view, we propose PatchGen, a text-free module that learns a sample-dependent soft predictive-subset mask as a task-driven proxy for the unobserved oracle subset mask. Specifically, histopathology visualizations suggest that PatchGen assigns higher scores to tumor-consistent regions than to some frequently co-occurring inflammatory context. Extensive experiments on natural and histopathological image
benchmarks spanning all three shift settings show that PatchGen improves average performance over matched-backbone baselines in most evaluated configurations, enhances generalization to unknown classes, and remains competitive with
vision-
language methods without text supervision.
Read the full abstract ▾
Machine Learning
Beyond Outcome Rewards: Step-Level Self-Distilled Policy Optimization for Deep Search Agents
Haoze Wu, Chuqiao Kuang, Tianyi Zhuang et al. (+1 more)
Deep search
agents operate over trajectories spanning dozens of steps, yet standard reinforcement learning provides only a single outcome reward per trajectory, which is far too sparse for effective credit assignment. On-policy self-
distillation (OPSD) addresses this by using the model's own logits as dense
token-level teachers, but extending it to search
agents introduces a fundamental tension: the teacher, having access to privileged information such as the correct answer, produces a distribution that differs systematically from the student's exploration-based
reasoning, and naive
distillation causes the student to inherit this information asymmetry rather than learn better search strategies. We resolve this tension through two contributions. First, we construct Evidence Anchors, which are concise, step-level evidence snippets extracted from the web, as privileged information that captures key
reasoning steps without revealing the entire answer path. Second, we propose Step-Level Self-Distilled Policy Optimization (SSPO), which converts teacher-student disagreement into step-level advantage weights within GRPO, applied exclusively to incorrect trajectories. This design decouples what to update from how much to update: the outcome reward determines the direction of policy change, while the teacher modulates its magnitude at each step. Correct trajectories are left untouched, preserving their diversity. On Qwen3-8B, SSPO consistently outperforms GRPO across BrowseComp,
GAIA, and FRAMES, surpassing or matching GRPO trained with twice as many gradient steps while adding only about 5 percent overhead per step from a single additional forward pass.
Read the full abstract ▾
All authors Haoze Wu, Chuqiao Kuang, Tianyi Zhuang, Xiaoguang Li
Computational Engineering
ARIES-Mission2: A Zero-Shot Vision-Language-Action Framework for Fast Large-Scale Aerial Mission Generation
Junhao Wei, Yanxiao Li, Haochen Li et al. (+8 more)
Multimodal Large
Language Models (MLLMs) have shown strong semantic understanding capabilities, but their direct use in low-altitude Unmanned Aerial Vehicle (
UAV) mission generation remains limited by weak spatial optimization and inefficient route planning. To address this issue, we propose ARIES-Mission2, a zero-shot
Vision-
Language-Action (VLA) framework that decouples visual-semantic perception from physical route optimization. Given natural-
language instructions and satellite imagery, ARIES-Mission2 first uses DeepSeek-V3 for task parsing and Molmo-7B for zero-shot target grounding, and then converts detected pixel locations into GPS waypoints through geospatial interpolation. To reduce the redundant backtracking caused by raw VLM-generated visiting orders, the back end formulates multi-target
UAV traversal as a Traveling Salesperson Problem (TSP) and compares four candidate routes, including the raw VLM order and the routes optimized by PSO, GPSO, and IPSO. The minimum-cost closed-loop route is then selected for mission generation. Experiments on the
UAV-VLPA-nano-30
benchmark show that ARIES-Mission2 achieves a total flight distance of 62.43 km, reducing the route length by 21.6% compared with the unoptimized VLA baseline (79.66 km) and by 9.5% compared with manual human planning (69.00 km). The complete 30-task workflow takes 575.40 s, averaging 19.18 s per task, which is approximately 3.6 times faster than human expert planning. Component-level timing shows that VLM inference dominates the runtime with 19.02 s per task, while the TSP solver requires only 0.16 s per task. Scalability analysis further indicates that the TSP module maintains lower growth in computation time as the number of targets increases.
Read the full abstract ▾
All authors Junhao Wei, Yanxiao Li, Haochen Li, Yifu Zhao, Dexing Yao, Baili Lu, Zikun Li, Yapeng Wang, Sio-Kei Im, Dingcheng Yang, Xu Yang
Artificial Intelligence
PROVE-RT: Generating Mechanized Theorem Prover Scripts for Real-Time Systems using LLMs
Sadat Shahriyar, Shareef Ahmed and Abdullah Al Arafat
Schedulability analysis is essential for certifying real-time systems, but existing tests are often developed through pen-and-paper proofs that are difficult to scale, validate, and maintain. Mechanized verification in PROSA/ROCQ offers a rigorous alternative, yet manually constructing such proofs requires substantial domain expertise and proof-engineering effort. Recent successes of large
language models (LLMs) across a wide range of tasks make them promising candidates for generating PROSA/ROCQ scripts for mechanized
theorem provers. However, state-of-the-art LLMs often lack the PROSA-specific knowledge required to correctly use its modeling abstractions and proof patterns. This paper introduces PROVE-RT, an LLM-assisted framework for generating PROSA/ROCQ scripts to mechanize schedulability analyses in real-time systems literature. PROVE-RT guides generation through dependency-aware informal sketches,
retrieval from processed PROSA documentation, staged skeleton generation, and proof completion. We construct a mechanization-oriented corpus from 1, 191 real-time systems papers, containing 13, 134 informal sketches with dependency information. On a curated evaluation set, direct prompting of state-of-the-art LLMs fails to reliably generate valid PROSA mechanizations, whereas PROVE-RT achieves a success rate of 44.7%. These results show that
retrieval-guided and staged LLM assistance can improve automated mechanization of schedulability analysis in PROSA/ROCQ.
Read the full abstract ▾
Artificial Intelligence
Correct Is Not Governed: Provenance Integrity in Agentic Workflows
Jesus Salas
Agentic workflows are commonly evaluated by whether they reach the correct outcome. That is insufficient in institutional settings, where a correct action may rely on the wrong authority, an unsupported completion claim, or work made stale by a later change. We define governed execution as work whose decisions, completion, and response to change are supported by inspectable provenance. We present
Matrix, a deterministic causal-state layer that records authority and fact dependencies, verifies completion evidence, and selectively invalidates affected work. Across controlled comparisons, governed and direct workflows often reached the same outcomes, but only the governed path consistently preserved governing evidence, refused unsupported closure, and limited recovery to dependent tasks. A role-separated transfer challenge then failed: a deterministically enforced completeness contract severely over-blocked synthetic packets produced outside its authoring context. These results do not establish
Matrix as a general accuracy enhancer; they support its primary role as an institutional integrity layer for making agentic work auditable and independently verifiable.
Read the full abstract ▾
Robotics · Cross-listed from physics.space-ph
Genetic Fuzzy System-based Control for Final Approach of Spacecraft Rendezvous and Proximity Operations
Daegyun Choi, Donghoon Kim and Henzeh Leeghim
In-space servicing has been receiving great
attention to extend the operation of spacecraft with defective components. This requires rendezvous and proximity operations for a chaser to provide service to a target. This work constructs a fuzzy inference system-based controller for the chaser to reach the cooperative target on a circular orbit in the final approach phase while minimizing the energy consumption of the chaser. The offline training process performed by a genetic
algorithm deals with multiple initial relative positions of the chaser, and the trained controller is validated using a testing environment with disturbances, which differs from the training scenarios.
Read the full abstract ▾
Human-Computer Interaction
NavSight in the Wild: Understanding Real-World Use of a Mobile Augmented Reality Application for People with Low Vision in Outdoor Navigation
Yuheng Wu, Kexin Zhang, Ben Kosa et al. (+3 more)
The ability to navigate outdoors safely and independently is crucial yet challenging for people with low
vision (PLV). While various augmented reality (AR) systems for low
vision have been designed and evaluated in ideal lab environments, no research has investigated their real-world feasibility and challenges. We present NavSight, a mobile AR application that assists PLV in outdoor
navigation by recognizing important outdoor objects (e.g., curb, vehicle) and rendering real-time visual augmentations. Through a seven-day diary study with 12 PLV in real-world settings, we characterize the impact of NavSight on scene perception, users' configuration strategies on what objects to augment and how to augment them across scenarios, how users made sense of and responded to recognition errors, and the social acceptability of using NavSight in public. We further identify environmental factors affecting recognition, such as weather conditions, lighting and shadows, and nonstandard road markings and textures, as well as usability issues in daily use. We discuss these real-world challenges and derive design implications for future AI-powered assistive AR systems for outdoor use.
Read the full abstract ▾
All authors Yuheng Wu, Kexin Zhang, Ben Kosa, Ru Wang, Sanbrita Mondal, Yuhang Zhao