Named after the hundred-eyed watchman of Greek myth, Argus watches the education landscape: spotting new opportunities, pressure-testing the ventures we're building, and tracing every read back to the real-world signals behind it.
The evidence library: the raw signals the pipeline is watching across the education ecosystem. Every idea is built from these.
arXiv:2609.02396v1 Announce Type: new Abstract: Radiology reports are written primarily for clinicians, and their specialized terminology often makes them difficult for patients to interpret. As a result, many patients turn to publicly available Large Language Models (LLMs) to help explain their reports, despite well-documented risks of factual inaccuracies and hallucinations. Automated lay-summary generation has emerged as a promising alternative, yet the effectiveness of retrieval-enhanced and clinically informed approaches for radiology-specific communication remains underexplored. This study investigates the extent to which Retrieval-Augmented Generation (RAG) and Named Entity Recognition (NER) improve the quality, factual consistency, and readability of automatically generated lay summaries compared with standard LLM-based generation. We develop a framework combining NER-based extraction of clinically relevant findings with a RAG mechanism for contextual grounding, evaluated acros
arXiv:2609.02391v1 Announce Type: new Abstract: Political evasion refers to responses that engage with a question while withholding the requested information. Recent NLP work frames political evasion as a classification task using a two-level taxonomy of response clarity and fine-grained evasion strategies. Existing work on response clarity and evasion classification is limited to English, leaving open whether the taxonomy and model behavior transfer across languages and political contexts. We introduce PolERo, a dataset of 3,574 human-annotated question-answer pairs extracted from official transcripts of five Romanian presidents. We evaluate multiple classification approaches on both datasets under matched conditions, including TF-IDF baselines, fine-tuned encoder models, a proposed sliding-window encoder, and zero/few-shot LLM prompting. We study cross-lingual transfer through joint bilingual training and machine-translation-based data augmentation. Our results indicate that fine-tun
arXiv:2609.02366v1 Announce Type: new Abstract: Named Entity Recognition (NER) has achieved substantial progress since the advent of large language models (LLMs). Nevertheless, the recognition of long-tail and domain-specific entities remains challenging due to the deficiency in parametric knowledge. Retrieval-augmented generation (RAG) offers a promising remedy by injecting external knowledge, but it also introduces noise and unnecessary cost when dealing with familiar cases. In this paper, we propose NE-R1, a novel framework for adaptive retrieval-augmented NER. We design a "retrieval-on-demand" mechanism for NER. Then we integrate it into models by a two-stage training method: (1) multi-task instruction tuning initialization; (2) end-to-end RL optimization with CoT. To achieve reasonable selection between parameterized and external knowledge, we design a multi-dimensional reward considering both accuracy and retrieval benefit. NE-R1 achieves state-of-the-art performance on various b
arXiv:2609.02315v1 Announce Type: new Abstract: A single sentence often expresses multiple valid relational triplets, which makes Open Information Extraction (OpenIE) fundamentally a multi-output task. Existing neural systems handle this by autoregressive generation, which is flexible but slow and prone to redundancy, or by fixed-slot prediction, which is efficient but couples the extraction budget to training. We introduce DIFFIE which instead treats the stochasticity of conditional discrete diffusion as the extraction mechanism itself: independent reverse-diffusion trajectories over per-token role tags produce a pool of candidate triplets, which are clustered under lenient matching and ranked to form the output. Both the pool size and the number of returned extractions are inference-time choices, decoupling the extraction budget from training and exposing test-time compute as a tunable axis. DIFFIE achieves the new state of the art in CaRB (1-1) both F1 and AUC, and outperforms the s
arXiv:2609.02309v1 Announce Type: new Abstract: GUI agents increasingly operate across websites, mobile apps, and desktop environments, yet the field still reports progress primarily through task success. We argue that practical deployment depends equally on efficiency: how much context, computation, action budget, and runtime overhead an agent consumes while succeeding. This survey studies efficient GUI agents through an end-to-end systems lens that preserves the current technical axes of observation efficiency, context and memory efficiency, action efficiency, and planner-side/system efficiency. For each subsection, we expand the seed literature through targeted search plus backward and forward citation chaining, then synthesize the dominant mechanisms, reported efficiency signals, and new overheads they introduce. Across the literature, recent progress converges on a small set of recurring ideas: selective reading instead of full-context ingestion, global-to-local visual allocation,
arXiv:2609.02275v1 Announce Type: new Abstract: Large language models are trained to model conditional distributions over text, yet it remains inadequately understood whether they capture the full diversity of plausible outputs present in their training data. We study this question through an information-theoretic lens by comparing the conditional entropy of model-generated outputs with that of the corresponding training data. Given paired input-output samples, we use conditional entropy and its matrix-based analogue based on von Neumann entropy to measure output variability beyond what is explained by the conditioning input, without requiring multiple reference outputs for the same prompt. Across LLM families with publicly available training data, including OLMo, Pythia, and GPT-Neo, we consistently find that model-generated outputs exhibit lower conditional entropy than their training data, across different model scales, sequence lengths, and decoding strategies. We observe a similar
arXiv:2609.02272v1 Announce Type: new Abstract: Faithfully translating research papers into repository-level implementations remains challenging because papers often describe methods at a high level, leave implementation assumptions implicit, and require generated repositories to preserve method logic, evaluation protocols, and cross-file consistency. Despite recent advances in paper-to-code agents, their intermediate outputs are often presented as free-form plans or summaries that downstream coding agents may ignore, reinterpret, or compress, leading to algorithmic simplification and inconsistent repository structure. To address these challenges, we introduce PaperCompiler, a paper-to-code generation framework that compiles paper-grounded evidence into explicit repository-level implementation specifications. PaperCompiler grounds implementation-relevant evidence while preserving source provenance and distinguishing paper-supported, inferred, externally delegated, and unresolved inform
arXiv:2609.02172v1 Announce Type: new Abstract: Optimization-based jailbreak attacks such as Greedy Coordinate Gradient (GCG) achieve strong effectiveness and transferability by optimizing adversarial suffixes on white-box source models. However, existing GCG-based methods rely on averaged adversarial loss and deep greedy search, which can over-emphasize easy-to-jailbreak behaviors and overlook promising regions of the suffix space. We propose BOSS, a plug-and-play framework that improves GCG-based jailbreak optimization through breadth-oriented suffix search. BOSS uses Tail-Focused Adversarial Loss (TFAL), standard source loss, and behavior coverage to select terminal suffixes, then explores multiple short trajectories and selectively continues promising suffixes. Experiments on public benchmarks show that BOSS improves attack success rates across multiple GCG-based methods while reducing optimization time.
arXiv:2609.02163v1 Announce Type: new Abstract: Information-theoretic measures derived from autoregressive language models are widely used to characterize the expectations that shape human reading, but whether language-variety-specific training improves such psycholinguistic alignment remains unclear. This question is still open for Cantonese, where recent NLP evaluations reported mixed benefits from Cantonese-specific training relative to Mandarin-oriented or general-purpose models. Using naturalistic Cantonese eye-tracking data, we compare two within-family adaptation contrasts: CKIP GPT-2 Tiny versus its lightly Cantonese-adapted JED351 derivative, and Qwen2.5-7B versus CantoneseLLM-7B, which underwent substantially more extensive Cantonese continued pretraining and instruction tuning. From each model, we derive lexical surprisal, POS surprisal, entropy before the target, and entropy reduction. Lexical surprisal and the joint four-metric model consistently favor CantoneseLLM-7B, fol
arXiv:2609.02158v1 Announce Type: new Abstract: Legal Judgment Prediction (LJP) models are typically trained on documents that describe facts from a prosecutorial perspective. Existing datasets further exhibit severe label imbalance toward guilty outcomes. Consequently, these models suffer from "Guilty Bias", blindly accepting the prosecution's narrative as objective truth. Previous studies employing three-step reasoning structures or training on synthetically generated innocence data improve overall accuracy, but they still fail to mitigate bias at inference time. In this paper, we introduce OBJECTION, an inference-time pipeline that integrates an Adversarial Lawyer Agent into each 3-step reasoning of offense, unlawfulness, and culpability. Unlike generic critics, our agent actively challenges the model's presumptions of guilt by injecting legal defense arguments at each reasoning stage. To thoroughly evaluate this, we present a new "Natural Innocent" dataset including 3.4k real-world
arXiv:2609.02153v1 Announce Type: new Abstract: Grammatical error annotation in Chinese learner writing requires labels that are both consistent and linguistically meaningful. This paper proposes a layered scheme linking computational Chinese grammatical error correction (CGEC) with pedagogical error analysis. The scheme first identifies character- and punctuation-level orthographic errors, labeling them by edit operation and subtype. Other errors receive a three-layer core label combining edit operation, linguistic domain, and part of speech, with optional Chinese-specific extensions for aspect, modality, comparison, argument structure, and complements. Drawing on CGEC resources, learner-error taxonomies, and Mandarin grammar, the taxonomy is evaluated through a coverage analysis of automatically extracted MuCGEC edits and a preliminary consistency study in which five large language models apply it to a sample. The results support the layered approach while identifying category bounda
arXiv:2609.02131v1 Announce Type: new Abstract: Sentiment in social-media threads does not only vary across posts; it shifts as users react to claims, corrections, evidence, and hostility within a branching reply tree. We study why sentiment changes in rumor-centric conversation trees by treating discourse moves (e.g., denial/correction, evidence/link, toxicity/attack) as candidate interventions and asking (i) what sentiment a reply expresses, (ii) whether the sentiment shifts relative to its parent, and (iii) which prior message most plausibly drove the reply's sentiment. To support this setting, we introduce CaSiRe, a causal sentiment reasoning layer over public rumor conversation datasets that adds post-level sentiment labels, induced parent-child shift labels, calibrated multi-label intervention tags, and explicitly annotated causal-source labels. We then propose C$^{3}$T (Counterfactual Causal Conversation Transformer), a thread-structured temporal model that jointly predicts node
arXiv:2609.02115v1 Announce Type: new Abstract: Natural language interfaces to databases have traditionally suffered from three structural limitations: exclusive targeting of relational SQL, unconditional dependence on large language model (LLM) inference at query time, and absence of any runtime signal when generated queries are semantically incorrect. This paper presents text2ql, an open-source Python framework that addresses all three limitations through a language-agnostic Intermediate Representation (QueryIR) and a pluggable renderer architecture. A single seven-stage detection pipeline serves both SQL and GraphQL targets; a zero-LLM deterministic mode delivers 100% execution accuracy at a median latency of 3.2 ms with no API cost; and every generated query carries a runtime confidence score in [0.15, 0.97] computed from an additive signal model. Evaluated on 50-query random samples from the Spider and BIRD benchmarks (indicative results; full-set evaluation is planned), the LLM-b
arXiv:2609.02108v1 Announce Type: new Abstract: Diffusion language models (DLMs) have emerged as a promising alternative to the auto-regressive paradigm. With bidirectional attention and any-order generation, DLMs naturally fit infilling tasks, which require generating a middle span conditioned on both the prefix and the suffix. However, infilling is sensitive to the length of the span, while DLMs require the length to be fixed before generation. Although prior studies extend DLMs to dynamic lengths, they still suffer from two limitations. (i) Sensitivity to initial length. These methods require a preset length to initialize the search and are highly sensitive to this initial length, often yielding suboptimal results. (ii) Inference inefficiency. They either insert length-changing operations during generation or repeatedly search for an appropriate length using multi-step denoising confidence, both of which introduce substantial extra forward passes and computational cost. Therefore, w
arXiv:2609.02091v1 Announce Type: new Abstract: Knowledge editing provides an efficient way to update factual knowledge in large language models. However, malicious edits may introduce safety risks, making it necessary to reverse undesirable editing effects. Existing reversal methods for parameter-modifying edits mainly focus on global removal, which may also erase beneficial edits that should be preserved. In this paper, we study selective reversal of edited knowledge, where the goal is to reverse targeted edited facts while preserving the remaining edited facts. Based on the hypothesis that each edit is sparsely encoded within the dominant subspace of the edited matrix, we propose a spectral-based reversal framework that locates edit-sensitive components within the dominant singular subspace of edited weights. Experiments across multiple settings demonstrate the effectiveness of our method in reversing selected edits while preserving unrelated edited facts. These results suggest that
arXiv:2609.02089v1 Announce Type: new Abstract: Steering aligns large language models (LLMs) by injecting a bias into selected activations at inference time, offering a far cheaper alternative to weight-update methods such as supervised fine-tuning or reinforcement learning. However, most existing training-free steering methods are input-independent: a single direction is fitted once and shared across all inputs. This is fundamentally limiting as different inputs occupy different regions of the activation space and admit different optimal steering directions toward the same target concept, much as the gradient with respect to a fixed loss varies from input to input. We close this gap with IDEEA (Input-Dependent stEEring via Activation cluster matching), a training-free framework for input-dependent steering. IDEEA clusters the positive and negative activation supports per attention head, and solves an optimal-matching problem to construct a set of cluster-conditional directions, all ab
arXiv:2609.02056v1 Announce Type: new Abstract: Scientific knowledge graphs organize entities and relations extracted from scientific literature, but they remain inherently incomplete. Missing typed links in such graphs can therefore represent plausible scientific hypotheses, such as unexplored associations between materials and applications. However, scientific hypothesis discovery is challenging because true discoveries are extremely sparse among typed candidate pairs: graph neural networks (GNNs) are efficient but unreliable for ambiguous cases, while large language models (LLMs) are knowledgeable but too costly to apply exhaustively and are not naturally grounded in graph structures. We propose HyGRAIL, a cost-aware and evidence-grounded framework that combines heterogeneous GNN triage with LLM-based hypothesis review. HyGRAIL first uses a GNN to score candidate hypotheses and identify a validation-calibrated ambiguous region, routing only graph-uncertain cases to LLM review. For e
arXiv:2609.02054v1 Announce Type: new Abstract: Large Language Models (LLMs) are increasingly deployed in interactive systems where understanding user intent precisely is paramount. A key capability for such systems is effective question clarification, especially when user queries are ambiguous or underspecified. This paper introduces a novel tri-agent framework for the robust evaluation of an LLM's ability to engage in clarifying dialogue. Our framework comprises three distinct LLM-based agents: (1) a Question Clarifying Agent (QCA), the system under evaluation, tasked with identifying ambiguities and posing clarifying questions; (2) a Respondent Agent (RA), designed to simulate human user responses, potentially including irrelevant or challenging replies; and (3) an Evaluator Agent (EA), an LLM-as-a-judge, which assesses the quality of the dialogue based on a comprehensive set of metrics. We detail a methodology for synthetic data generation in the supply chain domain as an example.
arXiv:2609.02015v1 Announce Type: new Abstract: Instruction-tuning data are judged by quality metrics, and tuned models are judged by benchmarks, but both judgments pass through an output interface: the surface format in which an answer is written. Using gradient signatures across 12 tasks, four semantically equivalent interfaces, three model families, and controlled corruptions, we show that this interface confounds both measurements. Spectral statistics such as effective rank are provably invariant to interface rotation and empirically blind to semantic corruption, while the direction of the update carries the quality signal. The interface-varying residual is not noise: it identifies each unit's own target task perfectly across all three families. Capability itself is stored relative to the training interface: a skill that raises accuracy by more than 40 points under the training format can be nearly invisible under every other, and correcting a single generation budget flips the mea
arXiv:2609.01971v1 Announce Type: new Abstract: AI is rapidly advancing neuroscience, yet many laboratories fail to fully unleash its potential due to significant interdisciplinary barriers. While pre-trained neural models for physiological data are progressing quickly, their heterogeneous architectures and modality-specific constraints hinder systematic integration, selection, and evaluation. Despite recent advances in large language model (LLM)-based agent systems for intelligent scientific applications, existing approaches often still lack the domain expertise required to effectively select and coordinate diverse neuroscience pre-trained models and handle unique data types in this domain. We present NS-Copilot, an LLM-driven multi-agent system for neuroscience analysis that autonomously supports end-to-end workflows for diverse professional tasks. It unifies domain-specific pre-trained models and supports key neuroscience modalities, including EEG and extracellular spike data, throu
arXiv:2609.01936v1 Announce Type: new Abstract: A language model's prediction of its next token develops across layers, and lens methods track this process by decoding intermediate hidden states into tokens. But a lens reading reflects both the hidden state and the readout (the unembedding matrix) used to decode it. Many lenses are fit on a corpus, and we show that two lenses differing only in their fitting corpus can report different tokens for the same hidden states. We call this dependence corpus conditionality. To examine readout structure independently of the fitting corpus, we introduce Sparse Readout Prism (SRP), which decomposes the readout using only its weights and expresses any token logit or logit difference as a sum of contributions from sparse readout features. This reveals readout features as a new unit of analysis for lens readings, exposing structure that token identities can obscure and enabling comparisons across tokens, contexts, layers, and lenses. Replacing the or
arXiv:2609.01918v1 Announce Type: new Abstract: Energy poverty is nearly absent from NLP-for-social-good, and the little existing work is either static retrieval/QA or relies on carbon-intensive cloud LLMs, a self-defeating "computational irony" for a humanitarian setting. We present EqGrid, a closed-loop simulation in which a low-frequency, open-weight LLM policy agent sets price and carbon bounds and targeted subsidies over a community of empirically-grounded household personas, while high-frequency multi-agent RL traders clear a continuous double auction constrained by a physical distribution grid (IEEE-33-bus with Dynamic Operating Envelopes). Our contribution is threefold and directly addresses how to measure the social impact of AI: (i) grounded personas (region-matched socio-demographics) whose load curves are checked for shape and level realism against real smart-meter data; (ii) formal energy-poverty equity metrics (Energy Burden, Gini of EB, LIHC) showing the intervention red
arXiv:2609.01878v1 Announce Type: new Abstract: Activation steering suppresses undesired behaviors in language models by adding a steering vector to the hidden state during generation. Recent conditional methods such as CAST and DSAS improve the behavior-capability trade-off by deciding when to intervene, but once active, they apply the full dense vector to all hidden dimensions, regardless of whether a neuron carries concept information or already lies in the desired regime. We introduce dimension-level conditioning as a complementary axis of selectivity that also decides which neurons to intervene on. Our method, GAPS (Gated Activation steering via Posterior and Separability), combines two training-free gates: a static separability gate that restricts steering to neurons with statistically reliable concept information (via AUROC), and a dynamic posterior gate that steers a neuron only when its current activation is better explained by the undesired concept under a Gaussian model. The
arXiv:2609.01867v1 Announce Type: new Abstract: A major question in cognitive modeling concerns the behavioral alignment between large language models and humans across linguistic and non-linguistic tasks. Unlike standard LLMs, large reasoning models (LRMs) are optimized with reinforcement learning from verifiable rewards, encouraging correct solutions to reasoning tasks rather than preference-aligned responses. Recent work (de Varda et al., 2025) investigates the cost of thinking in humans and LRMs by comparing human reaction times with model reasoning traces across a range of reasoning tasks. We isolate this alignment by turning to abductive reasoning: unlike deductive tasks, its difficulty cannot be inferred from formal structure and offers no shortcuts a model could exploit to mimic effort without genuine search, providing firmer ground for empirical claims of shared effort. We find further evidence of alignment between LRM and human reasoning effort, as well as evidence that model
arXiv:2609.01833v1 Announce Type: new Abstract: Sentence-level recognition of depression symptoms is challenging because similar expressions can differ in symptom relevance, and language-model inference is insufficiently grounded in diagnostic definitions. This study proposes a two-stage framework separating symptom-candidate generation from definition-grounded verification. A contrastively fine-tuned sentence encoder generates a symptom candidate per sentence, and a fine-tuned language model verifies whether the candidate is present or absent using the sentence, its context, and a candidate-specific diagnostic definition, checking its judgment against that definition before answering. Evaluated against encoder, inference-based, medical, and general LLM baselines and a matched single-stage supervised classifier, the proposed pipeline attains the best accuracy and F1 scores of all methods, with rationales matching expert-authored annotations. A preliminary clinical audit indicates moder
arXiv:2609.01832v1 Announce Type: new Abstract: Patients with depression present with diverse symptom profiles, yet clinical practice routinely reduces this variation to a single severity score. Large language models (LLMs) can potentially capture various symptoms and their severity from patient speech. However, how depressive symptoms are represented inside LLMs remains poorly understood, limiting clinical trust. To examine whether internal model activations match clinician judgment, we analyzed the residual stream of Gemma-3-27B-PT using mechanistic interpretability techniques. Recording activations across symptom descriptions drawn from validated clinical instruments, we found that symptom groups geometrically separated the most at layer 21 across multiple distance metrics. Using Semantic Projection, we then projected held-out naturalistic text onto Symptom Vectors constructed from these instruments. The resulting per-symptom coefficients preserved clinician-annotated rank ordering
arXiv:2609.01828v1 Announce Type: new Abstract: Spoken dialogue state tracking recovers slot-value pairs from speech, where ASR errors concentrate in entity values and persist across turns, making it both a generation and an editing problem. A strong per-turn text editor corrects much of this but, operating on the transcript alone, leaves three recoverable errors: a value predicted inconsistently across turns, an omitted slot, and a value the audio does not support. We present AVERT, which scores each candidate value by combining cross-turn agreement with a trained audio-conditioned verifier and resolves the three error types with three operators, vote, add, and swap, each restricted to the slots where its error is common. On SpokenWOZ, a base speech-LLM reaches 33.04 JGA, a text editor 38.34, and AVERT 40.13, without retraining either. This is in the range of a 1B end-to-end system that consumes the full spoken history (39.32), though AVERT uses two 1B decoders rather than one. The au
arXiv:2609.01810v1 Announce Type: new Abstract: Farsi, spoken by more than 120 million people, lacks a comprehensive benchmark for dialogue generation and understanding. We introduce TALKFA, a unified benchmark comprising three complementary datasets: (1) WIKI-FADIAL, 4.2K Wikipedia-grounded dialogues for knowledge-grounded generation; (2) DAILYDIALOG-FA, 6.6K dialogues annotated for dialogue acts and emotions; and (3) PLAYDIAL-FA, 2.1K theatrical dialogues with sentiment labels. While LLMs assist data construction, every dialogue undergoes multi-stage review and revision by native Farsi speakers, and only the final human-approved dialogues are released. Experiments with six LLAMA and MISTRAL models show that LoRA substantially improves dialogue generation while requiring only 25-50% of the training data to recover over 90% of the final performance gains. Across classification tasks, FABERT achieves the best dialogue-act performance, LORA-MISTRAL-7B performs best on emotion recognition
arXiv:2609.01798v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly deployed on mobile devices, making energy efficiency a key deployment constraint, yet the energy impact of prompt design remains underexplored. This paper aims to understand how two prompt properties, cognitive load and phrasing pattern, shape the energy behavior of on-device LLM inference. We conduct a broad empirical study covering prompt properties, datasets, models, and devices, with phase-level profiling that separates prefill and decode energy. We find that cognitive load primarily affects the energy cost per token, while phrasing pattern affects energy largely through token usage. Our energy-quality analysis further shows that prompt design reshapes the attainable frontier differently across models, highlighting the need for model-aware prompt design in energy-efficient on-device LLM inference. Code, datasets, and scripts are available at https://amai-gsu.github.io/PromptProperty/.
arXiv:2609.01794v1 Announce Type: new Abstract: How do learners avoid overgeneralizations such as Tom laughed me without explicit negative evidence? Constructionists have posited two proposals that describe indirect negative evidence against overgeneralizations: preemption (which privileges exposure to near-synonymous construction---e.g., she made him laugh) vs. entrenchment (all exposures to a verb's grammatical usages, including cases like He laughed). We disentangle these hypotheses by running controlled rearing experiments on LMs trained on child-caregiver conversations, where we systematically remove preemptive vs. non-preemptive evidence. We find that while LMs avoid overgeneralizations, they do not show preemption at a verb-specific level, instead showing weak but non-zero evidence of abstract preemption. Combined with results from analyzing the LMs' training dynamics, we find that LMs treat competing structures as indirect positive---as opposed to negative---evidence in the ver
arXiv:2609.01788v1 Announce Type: new Abstract: Real-world communication often requires pragmatic reasoning: interpreting meanings implied through context and cultural convention rather than stated literally. Existing pragmatic evaluation remains largely limited to English and high-resource languages, leaving Indic languages unexplored despite their linguistic and cultural diversity. We introduce VakyArth, the first pragmatic benchmark for Indic languages, designed as a diagnostic evaluation covering Hindi, Punjabi, Tamil, and Malayalam. VakyArth evaluates models across five phenomena: deixis, speech acts, implicature, social pragmatics, and coherence; through multiple-choice questions, natural language inference, and translation, with all items authored by native speakers. Across multilingual large language models (LLMs) of varying families and sizes, we find consistent failures on pragmatic meanings rooted in Indic linguistic and cultural conventions. Our analysis shows systematic di
arXiv:2609.01772v1 Announce Type: new Abstract: Meme understanding goes beyond recognizing visual content or literal text; it requires implicit cultural knowledge and pragmatic inference that most vision-language models still lack. We introduce MemeCULT-1K, a multilingual benchmark of 1,000 South Asian memes in Bengali, English, and Hindi, where each meme is paired with a cultural context note and three human-written explanations, along with a supplementary set of 54 Bengali regional dialect memes. We evaluate thirteen popular Vision Language Models (VLMs) under two settings: meme-only and context-aware. Providing minimal cultural context yields consistent gains across all models and languages: mean SBERT similarity improves from 44.6 to 56.4 (+11.8), BLEURT from 37.3 to 42.3 (+5.0), and LLM-as-a-Judge scores from 2.57 to 3.43 out of 5 (+0.86). Fine-grained error analysis reveals that closed-source models fail mainly on entity and reference misidentification, while open-source models a
arXiv:2609.01737v1 Announce Type: new Abstract: Mobile payment applications in Nepal are graphically mediated and largely inaccessible to visually impaired users. This paper presents SpeakPay, a voice-first digital wallet, and documents the central technical contribution: a controlled study of domain adaptation for low-resource financial speech recognition. We introduce NepFinSpeech-403, a 403-utterance dataset of Nepali financial voice commands (send, load, and balance operations spanning 237 unique numerals), and fine-tune Whisper large-v2 with LoRA. On the held-out test set, the domain-adapted model reduces Word Error Rate from 129.95% (zero-shot baseline) to 42.58% --- a 67.2% relative reduction --- and improves Devanagari numeral recognition accuracy from 0.0% to 73.9%. We find that word-level metrics understate the practical task-level impact: domain adaptation improves the Transaction Success Rate from 1.67% to 33.33%, a roughly 20x gain. The improvement is consistent at the ind
arXiv:2609.01687v1 Announce Type: new Abstract: Grounded question answering systems should answer only when the supplied evidence supports the answer. In multi-hop QA, this requirement is difficult because partial evidence can make an unsupported answer appear plausible. We study selective answering through evidence sufficiency boundaries: for the same question, a model should abstain under unsupported or partially supported context, answer when the context first becomes sufficient, and keep the answer stable when redundant evidence is added. We introduce Evidence Sufficiency Boundary Training, a generation-native training framework that constructs ordered evidence chains and supervises the abstain-to-answer transition directly. The method combines level supervision, a boundary flip margin, post-boundary stability, and answer recall protection. We build evidence chains from HotpotQA, 2WikiMultiHopQA, and MuSiQue, then evaluate models with chain metrics, raw QA utility, and unsupported-
arXiv:2609.01658v1 Announce Type: new Abstract: Retrieval-Augmented Generation enhances Large Language Models by grounding responses in external knowledge, but multi-hop reasoning remains vulnerable to error propagation, where early retrieval failures confound subsequent steps. Standard outcome-based optimization only rewards the final answer, leaving intermediate retrieval and reasoning errors undetected. While existing process-based methods introduce step-level signals, they still score each step against the final answer, rewarding spurious successes where flawed retrieval coincidentally produces the correct answer. Step-level supervision in RAG requires evaluating both logical validity and evidential grounding at each step. We introduce PRO-STEP: we train a generative PRM that evaluates both dimensions, employ PRM-guided value tree search to construct preference pairs contrasting valid steps against flawed ones, and optimize the policy via step-level Direct Preference Optimization.
arXiv:2606.03557v2 Announce Type: replace-cross Abstract: As generative AI capabilities expand, AI-driven virtual worlds face a growing architectural challenge. Users interact through in-world interfaces in multimodal ways, yet their requests demand fundamentally different AI backend models and computational resources. Embedding these capabilities directly into virtual world systems reduces extensibility, complicates maintenance, and limits the ability to coordinate services distributed across edge and cloud infrastructure. This paper presents an SLM-based Agent Orchestration Gateway, a lightweight runtime coordination mechanism that decouples a virtual world client from heterogeneous AI backends through intent-driven service routing. An edge-deployed SLM classifies the semantic intent of each user prompt, a configurable service registry validates and resolves the routing decision, and the selected backend is invoked transparently, enabling new AI capabilities to be introduced in the v
arXiv:2505.21907v3 Announce Type: replace-cross Abstract: AI copilots represent a new generation of AI-powered systems designed to assist users, particularly knowledge workers and developers, in complex, context-rich tasks. As these systems become more embedded in daily workflows, personalization has emerged as a critical factor for improving usability, effectiveness, and user satisfaction. Central to this personalization is preference optimization: the system's ability to detect, interpret, and align with individual user preferences. While prior work in intelligent assistants and optimization algorithms is extensive, their intersection within AI copilots remains underexplored. This survey addresses that gap by examining how user preferences are operationalized in AI copilots. We investigate how preference signals are sourced, modeled across different interaction stages, and refined through feedback loops. Building on a comprehensive literature review, we define the concept of an AI co
arXiv:2605.09550v3 Announce Type: replace Abstract: Artificial intelligence is increasingly entering digital games through diverse functions. While prior work has shown that player attitudes toward game AI are strongly context-dependent, less is known about how these attitudes are structurally combined within different groups of players. This study addresses this gap by modeling players' cross-context AI acceptance as interpretable attitude profiles. Based on questionnaire data from 771 digital game players, we apply Archetypal Analysis (AA) to standardized acceptance ratings across eight representative AI application contexts in games. The analysis identifies seven distinctive profiles: AI-Skeptics, Broad AI-Supporters, Creative-Play Explorers, Experience-Oriented Supporters, Systemic Order Advocates, Emotion-Centered Supporters, and Governance-Skeptics. Exploratory one-vs-rest (OvR) logistic regressions further suggest that profile membership is associated with players' perceived AI
arXiv:2601.12718v3 Announce Type: replace Abstract: Information Problem Solving (IPS) is a critical competency for academic and professional success in education, work, and life. The advent of Generative Artificial Intelligence (GenAI), particularly tools like ChatGPT, has introduced new possibilities for supporting students in complex IPS tasks. However, empirical insights into how students engage with GenAI during IPS and how these tools can be effectively leveraged for learning remain limited. Moreover, differences in background--shaped by cultural and socioeconomic factors--pose additional challenges to the equitable integration of GenAI in educational contexts. To address this gap, we present an open-source dataset collected from 279 students at a public Australian university. The dataset was generated through students' use of FLoRA, a GenAI-powered educational platform that is widely adopted in the field of learning analytics. Within FLoRA, students interacted with an embedded Ge
arXiv:2506.02856v2 Announce Type: replace Abstract: Designing music-based affective technologies requires understanding of how perceptions of AI versus human authorship shape trust and authenticity. We investigate how listener perception of AI-generated versus human-composed music affects emotional resonance and regulation. Drawing on affective computing and human-computer interaction frameworks, participants listened to AI- and human-composed music across labeling conditions (Correct, Incorrect, or Unlabeled) and emotion cases (Calm and Upbeat). Participants rated preference, efficacy of target emotion elicitation, and emotional impact. Results showed participants found human-composed music more effective in eliciting their target affective states and linked humanness to imperfection, flow, and "soul," underscoring authenticity as central to appraisal and ultimately leading to design implications relevant to music-based HCI. These findings challenge the assumption that preference alon
arXiv:2609.02512v1 Announce Type: cross Abstract: Beauty assessments from Multimodal Large Language Models (MLLMs) are increasingly popular amongst users, companies, and aestheticians. This raises the question of whether these AI models can accurately reflect human judgments of attractiveness. In a pre- registered exploratory study, we compared the attractiveness ratings of 2,513 human participants to four widely used commercial AI models: Claude, Gemini, GPT, and Grok. Results showed that MLLMs systematically rate faces more favourably and within a narrower range than humans and, at the time of study, do not reproduce human ratings in absolute terms. However, MLLMs exhibit strong correlations with human attractiveness judgments, accurately tracking the rank-ordering of faces. MLLMs may judge faces by different cues than humans; only face age was a predictor of facial attractiveness in both humans and MLLMs, with inconsistent patterns across models for ethnicity and gender. AI models s
arXiv:2609.02510v1 Announce Type: cross Abstract: We study body-only, 12-class acted-emotion classification from skeleton motion under leave-performer-out (LPO) evaluation, a hard, underdetermined setting: chance is 8.3%, and a protocol-matched reproduced STGCN++ baseline reaches only 25.73 +/- 4.03% Macro-F1. We show that reliable gains come not from a new architecture but from combining eleven models with orthogonal error modes: under 10-fold LPO cross-validation on the labeled training performers, an equal-weight logit-mean ensemble reaches 36.80 +/- 4.00% per-fold Macro-F1, a protocol-matched +11.07 pp (+43% relative) over the same-split reproduced baseline. Our central contribution is a tested explanation suite: for a strong ensemble member, part-masking and counterfactual edits show (rather than assert) that its decisions depend on motion-grounded body-region evidence, and this region saliency aligns with rule-based Laban Movement Analysis (LMA) attributes far more than with clas
arXiv:2609.02376v1 Announce Type: cross Abstract: Acoustic sensing offers a promising non-intrusive approach for monitoring daily activities of older adults, yet speech privacy concerns remain a critical barrier to real-world deployment. We present a privacy firewall pipeline based on a U-Net encoder-decoder, trained entirely on synthetic data, that removes speech from ambient audio while preserving environmental sounds indicative of daily activities. Activity recognition is performed using VGGish transfer learning with an SVM classifier. Evaluated on the ESC-50 and SINS datasets across multiple speech content levels, the proposed model reduced residual speech to 0% VAD-detectable speech (Silero Voice Activity Detection) under all tested conditions, outperforming Facebook Denoiser (6.55% residual), SepFormer (36.34%) and ConvTasNet (47.21%) on ESC-50 at the 100\% speech level. On ESC-50 at 40% speech level, classification performance recovers to 85% precision and 85% recall after speec
arXiv:2609.02242v1 Announce Type: cross Abstract: AI assistants often collaborate by proposing candidate edits, plans, or designs that users evaluate before adoption. Existing assistance methods focus on proposal quality or user-goal inference, often assuming that the user can reliably evaluate any proposal, which can fail in practice because of bounded rationality. We study evaluability-aware proposal planning, where proposals serve both as task interventions and as probes for learning latent preferences and evaluation constraints, where the resulting belief updates then guide later proposals. We formalise this setting as ProSE, a hidden-parameter sequential assistance problem, and instantiate it with a KL-regularised bounded-rational binary response model in which acceptance trades off value gain against a distance-dependent evaluability penalty. Analysing the planning consequence of this likelihood reveals that likely accepted proposals and informative probes need not coincide, whic
arXiv:2609.01767v1 Announce Type: cross Abstract: Artificial intelligence (AI) is transforming brain-computer interfaces (BCIs) from task-specific neural decoders into adaptive systems that complete language, smooth movement, regulate rehabilitation support and adjust stimulation. These capabilities can increase speed, fluency, usability and clinical reach, yet conventional performance metrics may overlook losses in intent fidelity, authorship, agency, therapeutic challenge and durable clinical benefit. I define neuroadaptive overfitting as a closed-loop failure mode in which an AI-mediated BCI becomes over-optimized to short-term proxies of success, including reduced effort, rapid acceptance, lower workload or smooth task completion, while drifting from the user's durable goals. I then propose Slow-Fast BCI, a framework for pacing AI assistance according to decoder evidence, uncertainty, contextual and clinical stakes, fatigue, and user- or clinician-defined goals. The framework disti
arXiv:2609.02729v1 Announce Type: new Abstract: Occupants are a primary source of uncertainty in building energy consumption and management, yet existing occupant behavior models cannot capture adaptive and reasoning responses considering the occupant's personal history, current context, and the type of energy signal being delivered. This study presents BuildOcc, an open-source Python platform that grounds large language model agents in the American Time Use Survey (ATUS), a nationally representative diary dataset covering 16,684 respondents. Through BuildOcc, each simulated occupant agent can be instantiated with a demographic persona drawn from ATUS population statistics, a memory stream that accumulates and reflects on timestep-level observations, and an activity scheduler that samples empirically from ATUS time-at-activity distributions. The platform exposes a three-layer interface - Python library, REST API, and Model Context Protocol server - so that any building energy tool (Ene
arXiv:2609.02719v1 Announce Type: new Abstract: The objective of this study was to demonstrate the potential of generating eco-feedback that accounted for unique household contextual information, named as context-aware eco-feedback, through a large language model-integrated framework. Previous studies have introduced personalized eco-feedback, mostly relying on household energy use patterns; however, they frequently did not reflect distinct household characteristics, including their persona or non-negotiable routines, leaving eco-feedback ineffective and sometimes superficial. To address these limitations, we introduced a contextual engineering framework that generated eco-feedback using a self-consistency with chain-of-thought prompt that leveraged household energy analysis data, utility rate structures, and characteristic information. We conducted a rigorous empirical validation and a combinatorial evaluation analysis to assess this framework systematically. The former aimed to test
arXiv:2609.02611v1 Announce Type: new Abstract: Intelligent Tutoring Systems (ITSs) traditionally focus their adaptive support on cognitive aspects of learning. Although effective, little is known about how such systems can be enhanced by addressing students' emotional states. In particular, the role of mindful interventions for supporting student learning and experiences in adaptive math learning remains underexplored. We developed "Math with Matt", an ITS that leverages Large Language Models (LLMs) to provide both cognitive and emotional support in algebra learning. The system offers 1) an LLM-based mindful chat that delivers context-sensitive emotional support through a pedagogical agent Matt, and 2) mindful feedback and hint messages (not just evaluative) to enhance learning experiences and reduce math anxiety. We conducted a classroom study with 7th graders, comparing a Mindful version against a version with cognitive support only. Overall, the ITS reduced executive state-math anx
arXiv:2609.02582v1 Announce Type: new Abstract: Decoding visual perception from electroencephalography (EEG) is important for non-invasive brain-computer interfaces (BCIs). However, most existing visual decoding pipelines directly align EEG features with semantic features from pretrained vision models. Those EEG signals carry information at more than one level and this practice disregards the varying neural visibility of different visual components in EEG signals, leading to cross modal mismatches and incomplete information use. In this work, we address this limitation through layer-wise contrastive learning. For each subject, the intermediate CLIP layer that maximizes retrieval performance is selected as the Neural Visibility Optimal Layer (NVOL). Built on NVOL, a hierarchical framework couples retrieval and generation through a shared intermediate representation. The retrieval branch fuses multi-NVOL features, aligns them to image embeddings via contrastive learning, and applies cros
arXiv:2609.02436v1 Announce Type: new Abstract: Collaborative Brain-Computer Interfaces (cBCIs) offer a promising mechanism to augment team decision-making, but existing approaches rely exclusively on evidence available only after a decision has been made and reported, such as reaction time or stated confidence. This limits their use to explaining or discounting a decision after the fact, rather than informing a team's response before it is finalised. We tested whether spatial-covariance EEG features could instead provide a genuinely pre-emptive signal of an operator's decision correctness, available within the response window itself, and whether such a signal depends on cognitive workload. Using a continuous virtual reality target-detection task, participants (N = 23) completed a within-subject workload manipulation (High vs. Low). At the team level, weighting votes by this pre-emptive neural signal, available before a response is committed, produced substantial accuracy gains on cont