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:2607.26604v1 Announce Type: new Abstract: Knowledge-base construction and querying are typically optimized in isolation: retrieval-augmented agents operate over a fixed, externally maintained index, whereas construction receives no signal from downstream use. We present WikiLoop, a feedback-coupled framework that jointly learns to build and navigate an agent-native Wiki, a persistent linked-page knowledge base designed for machine navigation. A role-conditioned shared policy supports two interfaces: a Navigator retrieves evidence from the Wiki to answer queries, and a Builder proposes structured edits evaluated through downstream navigation. The Navigator follows a sufficiency-before-efficiency objective that applies retrieval-cost penalties only after full evidence has been collected. The Builder learns from utility differences: a frozen Navigator scores each candidate edit by its change in downstream performance, while a guard penalty discourages regressions on unrelated querie
arXiv:2607.26555v1 Announce Type: new Abstract: Multimodal fake news detectors often generalize poorly across domains because they learn to trust unreliable evidence: domain-specific shortcuts amplified by imbalanced data and semantically inconsistent text-image pairs that make cross-modal evidence unreliable. We propose Expert-Guided Mutual Distillation (EGMD), which learns what evidence to trust across the prediction pipeline. At the input level, input-level calibration encodes pair-level coherence as a shared gain before fusion. At the representation level, an expert-guided teacher aligns domain statistics and encourages domain-specific patterns to concentrate in specialized experts. At the decision level, prototype-anchored domain-specific students use mutual learning and dual-channel distillation to inherit the teacher's feature geometry and calibrated predictions while discouraging local domain priors. We further construct Weibo_Balanced, a domain-balanced benchmark that isolates
arXiv:2607.26497v1 Announce Type: new Abstract: Retrieval-augmented generation (RAG) methods range from lexical and dense retrieval to graph-based indexing and agentic search. They are usually evaluated on different benchmarks at one corpus size, leaving their accuracy-cost scaling unclear. To bridge this gap, we present a controlled corpus-scaling study of these four paradigms. A ladder of 28 strictly nested tiers grows from roughly 1,000 to 512,000 documents while questions and a fixed bedrock of relevant and adversarial documents remain unchanged. Under one reader and judging protocol, we measure official accuracy, construction and query tokens, and latency. Our experimental results show that BM25 scales best in this controlled setting: it defines the low-cost end of the Pareto frontier at every measured tier and leads accuracy from mid-scale onward, without LLM-based construction. The File-System Agent matches or slightly exceeds BM25 at the smallest tiers but uses 39 times more qu
arXiv:2607.26470v1 Announce Type: new Abstract: Multi-turn information-seeking conversations require both multi-hop reasoning and long-range dependency tracking across turns. However, existing RAG systems typically represent conversational memory as raw dialogue history, rewritten queries, or unstructured summaries, making it difficult to recover the specific prior reasoning steps and evidence required for follow-up queries. Our key insight is to align conversational memory with retrieval by representing dialogue context as sub-question-level reasoning traces. Building on this insight, we introduce MuMu-QA, a benchmark for multi-turn multi-hop RAG with explicit cross-turn sub-question dependency annotations, and CMT-RAG, a complementary memory framework for this setting. At each turn, CMT-RAG employs a state-space trace generator, whose recurrent state serves as runtime memory, to incorporate recent conversational context and decompose the current query into structured trace drafts con
arXiv:2607.26455v1 Announce Type: new Abstract: Large language models (LLMs) have demonstrated strong capabilities in knowledge acquisition and reasoning, yet their ability to retain previously acquired knowledge under repeated updates remains insufficiently understood. Existing evaluation paradigms primarily focus on single-step reasoning or static knowledge editing, which fail to capture the temporal dynamics of knowledge retention and degradation during continual model modification. In this work, we propose ForgetBench, a benchmark designed to systematically characterize forgetting behavior in LLMs under continual knowledge editing. ForgetBench introduces two complementary evaluation paradigms, namely concept-based QA and scenario-based QA, to disentangle isolated factual retention from structured relational knowledge preservation. Building upon a sequential editing framework, we construct temporally ordered knowledge streams and evaluate model behavior across multiple editing stage
arXiv:2607.26448v1 Announce Type: new Abstract: A known limitation of long-context language models is their increasingly unreliable performance in non-additive, set-based aggregation as context length grows. Examples include cardinality estimation, set relationships, and grouped statistics, which widely exist in logs, program outputs, tables, and multi-turn conversations. To provide the aggregation state required by these tasks, we introduce a model-side aggregation interface that maintains compact Hash-based HyperLogLog (HLL) sketch states alongside a frozen language model. While the model processes the context, an extractor maps each relevant record to a canonical identity. The identity is then hashed and updates the HLL state. These states can be merged across context segments and/or read out directly for downstream reasoning, avoiding an additional generate-execute-return cycle. We validate the proposed approach by setting the HLL state size as 2 KiB (2,048 registers), which does n
arXiv:2607.26410v1 Announce Type: new Abstract: We present Voice Memory, a inference-only scheme for agentic speech recognition: at stream time, a frozen corrector reads a single per-domain memory.md and decides per utterance whether to act on the hypothesis or abstain and keep the 1-best. Asynchronously, a score-gated optimizer revises that file through bounded edits, accepting an edit only when it strictly improves a held-out score. Extended from classical ASR-LM framework, we refer this split the listener-thinker architecture; the two roles are coupled only through the memory, so no weights change and the learned skill stays auditable and portable. Restraint turns out to be the operative skill this loop discovers: unconstrained generative error correction (GER) over-corrects, breaking correct tokens on up to 64% of its edits on financial news, and Voice Memory, reduces this rate to 35%. Across ten HyPoradise domains with an open corrector, Voice Memory, lowers weighted word error ra
arXiv:2607.26397v1 Announce Type: new Abstract: Enzyme function prediction is a hierarchical, knowledge-intensive form of protein function classification. Existing benchmarks expose an anomaly: general LLMs often get the coarse first level right, yet once asked for a complete EC number their accuracy at levels two through four drops to almost zero, while specialized models and tools stay usable. We propose EC-Reason-Bench, a training-free, diagnostic evaluation protocol built to answer two questions: why general LLMs score close to nothing on EC number prediction, and how much of that loss can be recovered without updating a single weight. We break enzyme classification ability into four orthogonal levers that can each be measured on their own: output structure, external knowledge, reasoning structure, and reasoning robustness. We test each lever with an inference-time method against a shared zero-shot baseline reproducing previously reported near-zero performance. Experiments with sev
arXiv:2607.26389v1 Announce Type: new Abstract: Fine-tuning a language model on data containing a narrow flaw, such as insecure code or incorrect mathematical answers, can cause broad misalignment through a mechanism that remains debated. We provide an interpretable account: in the models and corpora we study, misalignment behaves like a shift in personality. Prior work extracts activation directions for character traits from a single binary contrast, which can separate or steer behavior without establishing a calibrated scale. We instead extract personality vectors for the Big Five using a graded, three-level intervention and validate them on two open-weight models. The three levels are linearly ordered, with Cohen's d values of up to 6.2; the vectors transfer zero-shot and trait-specifically to an independent corpus; and their effects are strongest within a middle-layer band. Applied to training data, the vectors reveal that misaligned corpora across eight domains share a common Big
arXiv:2607.26368v1 Announce Type: new Abstract: Financial disclosures contain numerical claims, temporal statements, entity references, policy commitments, and risk descriptions that may conflict in qualitatively different ways. Detecting a conflict is only the first step: review workflows may also need to determine its type, since numerical, temporal, referential, factual, and normative inconsistencies require different evidence and downstream checks. We study this problem as fine-grained inconsistency classification. Using a fixed 5,940-instance snapshot of SBID-FD, a synthetic financial-disclosure benchmark with 11 inconsistency labels and paired reference evidence spans, we compare frozen embedding classifiers, fine-tuned encoders, evidence-augmented classifiers, prompted large language models, and LoRA-adapted generative models under a shared evaluation protocol. A fine-tuned 300M encoder reaches 61.9% accuracy, compared with 61.5% for a LoRA-adapted Qwen3.5-9B model and 61.3% for
arXiv:2607.26355v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly embedded in everyday life and widely used for information seeking, raising concerns about their potential to perpetuate social biases and reinforce stereotypes. In this study, we investigate gender bias in LLMs through the lens of their associations with musical instruments. Building on social-science research on the cultural gender-typing of instruments, we introduce Symphony-Bias, a parallel multimodal dataset spanning text, vision, and audio. We evaluate ten multimodal models with diverse architectures and scales across 22 musical instruments, analyzing how they associate each instrument with three gender categories: {male, female, non-binary}, across three modalities: {text, vision, audio}. Our results show that 92\% of instrument-level outcomes align with prior social-science findings, with the harp and drums showing particularly consistent gendered associations across all evaluated model
arXiv:2607.26286v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly used as general-purpose translation systems, but their behavior is usually evaluated under a single prompt shape: translate one source sentence into one target language. In practice, users may ask for one target language, for several related languages at once, or for translations conditioned on examples. This paper studies prompt scope and demonstration selection as experimental variables for local LLM machine translation. We evaluate English-to-Romance and English-to-Germanic translation on the full FLORES devtest split for nine official European Union languages. We compare three local instruction-tuned LLMs, llama3.2:3b, mistral:latest, and qwen2.5:14b, against dedicated MT baselines from OPUS-MT and NLLB-200. We test zero-shot prompting and k=5 few-shot prompting with random, lexical-similarity, and embedding-similarity demonstration selection. We also compare single-target prompts with JSO
arXiv:2607.26250v1 Announce Type: new Abstract: Line-haul trucking costs are dominated by three comparably sized components: energy, driver labor, and equipment. Most efficiency technologies address only one component at a time. This paper presents Robostreet Flow, a freight architecture that jointly optimizes the vehicle, convoy formation, and operating model to minimize cost per ton-mile on high-volume point-to-point corridors. The Flow platform is a battery-electric 6x4 tractor with a teardrop single-seat cab and a drag coefficient of 0.35, approximately 40% below that of conventional Class 8 tractors. A carbon-composite monocoque and structurally integrated batteries reduce net vehicle weight by 50%. A 513 kWh tractor battery and a 340 kWh powered trailer battery provide a 500-mile single-charge range. Four Flow trucks operate as a coordinated convoy with a safety driver only in the lead vehicle, while three followers operate in SAE Level 4 automated mode. Computational fluid dynam
arXiv:2607.26228v1 Announce Type: new Abstract: Instruction hierarchies are a core safety assumption of language model deployment: higher priority inputs, such as system prompts, should override conflicting lower priority inputs from users or tools. Yet frontier LLMs often violate this hierarchy. We introduce V-Steer, a training-free inference time method that restores privileged influence by editing cached value vectors at prompt positions. Using direct logit attribution on the first next token prediction, V-Steer identifies heads where lower priority spans dominate privileged ones, then boosts privileged spans and suppresses conflicting lower priority spans through in-place multiplicative edits to cached V tensors. Since the method acts only on cached values, it remains compatible with fused attention backends and adds only a one time prefill overhead. Across models from 7B to 70B, this attribution guided intervention raises primary constraint accuracy from under 18% up to 92% on con
arXiv:2607.26221v1 Announce Type: new Abstract: With the advancement of AI technologies, Generative AI (GenAI) and human written text have become nearly indistinguishable. Additionally, the global standardization of AI chatbots made academic malpractice more frequent. Furthermore, existing research indicates GenAI poems are the most difficult to distinguish even without any modification thus, GenAI poems are naturally deemed human-like by modern detectors. However, the objectivity of such dissertations needs to be verified against modern detection tools but the subjectivity of poetry and the black-box nature of the modern LLMs (Large Language Models) architectures made verification of such work quite complicated. Hence, the main objective of the research is to deduce the attributes of English poetry that contribute classification and misclassification of both human and AI poems and provide corroborating or contradicting evidence to the poetry distinguishability claim. For such characte
arXiv:2607.26200v1 Announce Type: new Abstract: Content-moderation classifiers are usually evaluated in isolation, but deployment requires choosing where to intervene and what follows a flag. We evaluate these choices using two end-to-end customer-outcome metrics rather than component accuracy: Usefulness, the fraction of turns with a shown, non-harmful, relevant response, and Harmful Exposure, the fraction with a shown harmful response. Latency and error rates are diagnostics. We compare Input only, Response only, and Input + response hard blocking on a human-labelled product benchmark and public ToxicChat evaluation. At the evaluated operating points, Response only achieves the highest filter-only Usefulness in both settings, while Input + response achieves lower Harmful Exposure. Replacing Response only blocking with Response + rewrite recovers most blocked traffic and yields the same observed Harmful Exposure count as Response only blocking for the selected configuration; this equa
arXiv:2607.26178v1 Announce Type: new Abstract: Turn-taking is a central component of full-duplex interaction. Which turn-taking behaviors are appropriate varies with the scenario, yet current models apply a single norm regardless of context. This limitation originates in their training data: human-human speech corpora capture natural timing phenomena but provide little role grounding or scenario-specific norms, while heuristic or prompted synthesis methods inject turn-taking behaviors without basing them on human preferences. We introduce DuplexGen, a framework for generating dialogues with scenario-adaptive turn-taking by calibrating LLM predictions against a small set of slot-level human preference annotations. In six cooperative and competitive tasks, human turn-taking preferences differ systematically, and DuplexGen aligns substantially more closely with those preferences than uncalibrated prompting or training solely on generic human-human data; a full-duplex model trained on Dup
arXiv:2607.26066v1 Announce Type: new Abstract: The growing volume of scientific submissions has motivated interest in using large language models (LLMs) to assist peer review. Existing automated novelty assessment approaches typically compare a paper's claimed contributions against prior literature, implicitly assuming that these contributions are accurately realized in the work itself. Human reviewers, however, frequently challenge novelty claims not because similar ideas already exist, but because the methodological evidence presented in the paper does not adequately support them. This internal mismatch between claimed contributions and methodological realization is rarely examined by current LLM-based review systems. To address this gap, we introduce intra-paper claim verification, a framework that evaluates whether novelty claims articulated in a paper are substantiated by the methods used to realize them. The framework employs an LLM to extract novelty claims from the introductio
arXiv:2607.26060v1 Announce Type: new Abstract: LLM-based chatbots are transforming customer service in regulated domains such as banking, but scalable and cost-effective validation remains a critical barrier to safe deployment. We present a two-part contribution for large-scale chatbot validation. First, we introduce a methodology for creating high-fidelity synthetic customer agents (SCAs) as digital twins, grounded in real transactional and conversational data, that enables automatic generation and behavioral conditioning to simulate diverse customer profiles and interaction styles. Evaluation demonstrates that SCAs achieve high semantic alignment with real customers, low hallucination rates, and successful personality trait reproduction with controllable interventions. Second, we develop an SCA-based validation framework combining automated LLM-as-a-Judge evaluation, human expert testing, and adversarial probing. Scenario-based validation across emotional states, demographic groups,
arXiv:2605.30256v2 Announce Type: replace-cross Abstract: Natural human conversation is full-duplex and audio-visual: people simultaneously speak and listen while continuously interpreting and producing nonverbal cues, such as nods, smiles, and gestures. To support successful human-agent interaction, agents must model full-duplex audiovisual conversation; however, existing full-duplex benchmarks evaluate only speech. In this work, we present VideoFDB, the first benchmark to evaluate full-duplex audio-visual-to-audio-visual (AV2AV) conversational agents. VideoFDB contributes (i) 237 dyadic clips spanning 11 nonverbal conversational dynamics from real-world video calls, (ii) a taxonomy separating perception from generation behaviors, and (iii) a rubric-based LM-as-judge evaluation framework with interpretable axes for assessing conversational quality with respect to nonverbal conversational dynamics. Across open- and closed-source vision-speech agents, we find systematic failure modes: c
arXiv:2509.07260v5 Announce Type: replace-cross Abstract: Mobile and wearable healthcare monitoring play a vital role in facilitating timely interventions, managing chronic health conditions, and ultimately improving individuals' quality of life. Previous studies on large language models (LLMs) have highlighted their impressive generalization abilities and effectiveness in healthcare prediction tasks. However, most LLM-based healthcare solutions are cloud-based, which raises significant privacy concerns and results in increased memory usage and latency. To address these challenges, there is growing interest in compact models, Small Language Models (SLMs), which are lightweight and designed to run locally and efficiently on mobile and wearable devices. Nevertheless, how well these models perform in healthcare prediction remains largely unexplored. We systematically evaluated SLMs on health prediction tasks using zero-shot, few-shot, and instruction fine-tuning approaches, and deployed t
arXiv:2504.20903v4 Announce Type: replace-cross Abstract: How should organizations divide and sequence decision tasks between human and artificial agents? We develop a computational model of joint sequential adaptation in which two agents differ in a single, precisely specified way: the memory regime governing how past decisions shape subsequent ones. A recency-weighted regime, motivated by behavioral evidence on human adaptation, privileges recent outcomes; a uniform-memory regime, motivated by the scale-free consistency of algorithmic updating, weights a window of past outcomes equally. Situated in the lineage of NK/NKC models but developed on its own terms as a sequential-adaptation model, the framework varies task scope (N), within-task coupling (K), and cross-agent coupling (C) across modular and sequenced task structures. Three mechanisms organize the results. First, threshold dynamics create absorbing high- and low-payoff regimes, so adaptation compounds whatever it inherits. Se
arXiv:2604.20781v2 Announce Type: replace Abstract: This paper reports on the process of designing the UK Co-Benefits Atlas, which communicates and publicizes data for climate mitigation. Visualization atlases--an emerging type of platform to make data about complex topics comprehensive through interactive visualizations and explanatory content--pose challenges beyond traditional visualization projects. Atlases must address diverse and often uncertain audiences and use cases, support both explanatory and guided exploration, and accommodate complex, evolving data. Over 10 months, our team of visualization and domain experts conducted 8 design workshops, iterative prototyping, 15 stakeholder onboarding sessions, and continuous reflection. These intertwined processes informed the development of the Atlas, comprising over 400 pages of visualizations and explanations. They also enabled a deeper understanding of how stakeholders may critically engage with the atlas in practice, in terms of i
arXiv:2604.15336v2 Announce Type: replace Abstract: Large language models (LLMs) enable increasingly capable tutoring-style conversational agents, yet effective tutoring requires sensitivity to learners' affective and cognitive states beyond text alone. Facial expressions provide immediate and practical cues of confusion, frustration, or engagement, but remain underexplored in LLM-driven tutoring. We investigate whether facial-expression-aware signals can improve empathetic tutoring responses through prompt-level integration, without end-to-end retraining. We build a scalable simulated tutoring environment where a student agent exhibits diverse facial behaviors from a large unlabeled human facial expression video dataset, and compare four tutor variants: a text-only LLM baseline, a multimodal baseline using a random facial frame, and two Action Unit estimation model (AUM)-based methods that either inject textual AU descriptions or select a peak-expression frame for visual grounding. Ac
arXiv:2602.23971v4 Announce Type: replace Abstract: Sycophancy, the tendency of large language models to favour user-affirming responses over critical engagement, has been identified as an alignment failure, particularly in high-stakes advisory and social contexts. While prior work has documented conversational features correlated with sycophancy, we lack a systematic understanding of what provokes or prevents AI sycophancy. Here, we present a set of controlled experimental studies where we first isolate how input framing influences sycophancy, and second, leverage these findings to develop mitigation strategies. In a nested factorial design, we compare questions to various non-questions where we vary three orthogonal factors: epistemic certainty (statement, belief, conviction), perspective (I- vs user-perspective), and affirmation vs negation. Measuring expressed sycophancy, how sycophantically a model phrases its free-text response, we show that (1) sycophancy is substantially higher
arXiv:2511.10026v3 Announce Type: replace Abstract: Providing haptic feedback via smartphone touch screen may potentially offer blind people a capability to understand graphs. This study investigated the discrimination performance of haptic gratings in different frequencies, in both visually impaired (VI) and sighted (S) individuals. 6 VI participants and 10 S participants took part in two experiments designed to compare their ability to interpret grating images with a finger swiping across a smartphone touchscreen without vision. The swipe gesture activates phone vibration temporally synchronized with the black stripes. Their tasks were: (1) determining whether a grating pattern is presented on the touchscreen, (2) comparing two different grating frequencies and determining the wider one. Results demonstrated that the VI group exhibited superior tactile sensitivity compared to the S group, as evidenced by their significantly better performance in Experiment 1 (accuracy of 99.15\% vs.
arXiv:2505.10300v2 Announce Type: replace Abstract: Responsible AI (RAI) efforts increasingly emphasize the importance of addressing potential harms early in the AI development lifecycle through social-technical lenses. However, in cross-functional industry teams, this work is often stalled by a persistent coordination challenge: how technical roles hand off technical intent, how teams establish shared structures for collaboration, and how non-technical roles are supported in systematically evaluating harms. Through literature review and a semi-structured interview study with 8 practitioners, we unpack how this challenge manifests---technical design choices are rarely handed off in ways that support meaningful engagement by non-technical roles; collaborative workflows lack shared, visual structures to support mutual understanding; and non-technical practitioners are left without scaffolds for systematic harm evaluation. Existing tools like JIRA or Google Docs, while useful for product
arXiv:2607.27189v1 Announce Type: cross Abstract: We introduce APEX-Accounting, a benchmark built by Mercor in partnership with Ramp, to assess whether frontier models can do the real work of accountants. Tasks include reconciling accounts, accruing expenses, posting transactions, and producing reports. The private eval set comprises 160 tasks, split across 10 worlds. Each world contains an accounting system, as well as spreadsheets, PDFs, and other files. Every task was authored and solved by experts in accounting and bookkeeping, who also wrote grading rubrics. Across nine frontier models, Claude-Fable-5 (Max) leads with 56.4% Mean Criteria@3, ahead of Muse-Spark-1.1 (xHigh) at 52.6%. No model scores more than 2.6% Pass^8 (GPT-5.6-Sol (Max+Pro)) and the highest Pass@8 is 21.5% (Muse-Spark-1.1 (xHigh)). We experiment with increasing the token budget from $1 to $50 and observe an instance of Simpson's paradox: scores increase as the token budget increases but within a given budget-cons
arXiv:2607.27177v1 Announce Type: cross Abstract: Effective collaboration with novel and diverse partners is a crucial skill for autonomous agents. Most current ad-hoc teamwork (AHT) approaches assume that agents will collaborate on a single, fixed task and that the partner's capabilities, their ability to successfully execute the desired action, are already known. In reality, a partner's true capabilities are often hidden, and human collaborators may act sub-optimally on tasks with multiple valid strategies. To address these limitations, we extend ad-hoc teamwork into a multi-task setting by re-framing it as a problem of joint planning with decentralised execution under hidden partner capabilities. We introduce CE-CM (Capability Estimation via Contextual Models), an approximate Bayesian method that infers task-invariant capability vectors. By using simulation-based sampling, the agent estimates capabilities and induces a contextual Multi-agent Markov Decision Processes for planning. T
arXiv:2607.27155v1 Announce Type: cross Abstract: Large language model (LLM) agents are increasingly expected to assist users in completing tasks. However, existing benchmarks provide limited support for evaluating whether agents can carry out office-suite workflows at a reasonable cost. We introduce OmegaUse-OfficeVal, a benchmark for evaluating LLM agents on long-horizon office-suite tasks with task-level economic grounding. The benchmark comprises 100 tasks derived from office-suite requests proposed by practitioners and adapted through a privacy-preserving process. On average, these tasks require 2.32 hours of human labor to complete. An important feature of the benchmark is that each task is paired with two economic signals: human labor time and task price proxy. These signals enable direct comparisons between human costs and LLM inference costs, as well as value-weighted evaluation. To support stable evaluation, we develop code-based verifiers from fine-grained rubrics. We evalua
arXiv:2607.26746v1 Announce Type: cross Abstract: Accurate identification of Alzheimers disease (AD) using resting-state functional magnetic resonance imaging (rs-fMRI) remains challenging due to the high dimensionality, noise, and complex inter-regional dependencies inherent in functional brain connectivity, which limit the effectiveness of traditional approaches based on handcrafted connectivity features or conventional machine learning models. In this work, we present an attention-based deep learning framework for Alzheimers disease classification that operates directly on rs-fMRI functional connectivity matrices by treating brain regions as tokens and employing a Transformer-inspired self-attention mechanism to model long-range and global functional dependencies across distributed brain networks. The proposed framework learns discriminative functional representations without reliance on manual feature engineering and is evaluated on a longitudinal cohort from the Alzheimers Disease
arXiv:2607.26640v1 Announce Type: cross Abstract: Current human evaluation of machine translation typically assesses single outputs in isolation, a paradigm that suffers from high annotator noise and cost. We introduce Contrastive Error Span Annotation (cESA), a protocol that presents multiple translations of the source input (text, video, audio, image). In cESA, the annotator sees multiple translations of the same document, marks major and minor error spans, and then assigns a score from 0% to 100% on absolute scale. By allowing annotators to access the shared context across multiple outputs, cESA facilitates more consistent and efficient judgments. We validate cESA using a large-scale human evaluation of English->Japanese translations of 12 models, demonstrating reductions in annotation time and noise compared to standard pointwise evaluation. Unlike existing contrastive ranking methods, cESA yields absolute quality judgments that enable simple, interpretable non-parametric model ran
arXiv:2607.26611v1 Announce Type: cross Abstract: AI-assisted coding increasingly translates informal user intent into executable software, yet coding requests often contain ambiguities that recur in user-specific ways across tasks and sessions. Existing disambiguation methods typically address each ambiguous request in isolation within the current coding session, often through eliciting additional clarification. However, whether resolved session history from the same user can serve as memory for resolving recurring personalized ambiguity in a newly opened session remains underexplored. We formulate personalized ambiguity adaptation as a new task: given a user's previously resolved coding sessions and a new ambiguous request, an assistant should identify the recurring ambiguity pattern, produce the intended executable solution, and minimize clarification. To benchmark this task, we introduce CAPA, which characterizes personalized coding ambiguity through six mechanisms and injects thes
arXiv:2607.26375v1 Announce Type: cross Abstract: Coding agents (e.g., Cursor) improve developer productivity by optimizing task completion, but shifting users from writing code to prompting and reviewing may harm their understanding, impeding oversight, learning, and communication. To probe this, we have 54 students create a website with one of two AI systems: an agent that edits user code; or a chatbot where users write code alone or adapt generic code snippets. We test understanding via comprehension questions and a task where users extend their code without agents, showing: (1) While agents aid initial task completion, they harm users' code comprehension and thus do not prepare users to extend their code; (2) Low-effort agent interaction types, like copy+paste prompts and auto-accepted edits, are linked with lower comprehension; and (3) Despite self-reported weaker understanding, users still prefer coding agents because they are quick and easy to use. While users stay in the loop f
arXiv:2607.26300v1 Announce Type: cross Abstract: AI agents are increasingly adept at tackling complex, long-running tasks. With the rapid surge of autonomous capabilities, human oversight is systematically lagging behind due to limited human-centered interfacing. Aiming to address this, we introduce AgentGUI, a user-friendly, locally hosted GUI for seamlessly observing and steering AI agents amid multiple concurrent, long-running sessions. AgentGUI features 1) rich agent trajectory visualizations, 2) effective manual and automated steering, and 3) integration with and coordination between open-source and frontier agent frameworks. A controlled user study demonstrates statistically significant reduction in the time it takes to identify key elements from agent traces (38% faster, p = 0.023). In a preliminary experiment, AgentGUI's automated drift prevention feature raises the task completion rate of small local agents by as high as 34pp across a 0.8B--9B model ladder (N=50 runs per mode
arXiv:2607.26278v1 Announce Type: cross Abstract: It is common for two-dimensional embeddings of high-dimensional data to be read far beyond what they can support. Distances in and between clusters, the meaning behind empty spaces, and the amount of structure hidden at each point are generally invisible in the output of methods such as t-SNE and UMAP. This is because the information that could support the meaning of these properties is discarded during the optimisation process. Here, we present FloDR, a dimensionality reduction method that embeds data through an invertible normalising flow. While FloDR only uses the first two output coordinates to create a two-dimensional embedding, it retains the remaining coordinates rather than discarding them. In addition to the embedding, an exact inverse and an exact density are properties of a trained mapping, which enable diagnostic visualisations that are computed from the exact inverse of the model that drew the layout rather than from an app
arXiv:2607.26109v1 Announce Type: cross Abstract: Why do some teams consistently mobilize collective effort and achieve superior performance while others struggle to coordinate action? We introduce Attention-Directing Ability (ADA), a latent team capability capturing how effectively members' interaction signals elicit engagement and coordinated responses from others. Extending the attention-based view, we conceptualize attention direction as an emergent coordination capability embedded in patterns of interaction rather than as a cognitive state or an outcome. Teams differ in the extent to which attention-directing signals trigger collective responses, and these differences shape how teams mobilize effort and perform. We examine ADA in a distributed innovation effort involving 2,233 participants collaborating asynchronously in 79 self-organized teams across 165 public Slack channels, generating over 30,000 messages. We model the causal responsiveness among interaction signals and derive
arXiv:2607.26074v1 Announce Type: cross Abstract: Reproducibility has become a cornerstone of credible recommender systems research, driven by growing concerns about the reliability and generalizability of experimental results. In response, the ACM RecSys conference introduced a dedicated Reproducibility Track in 2020 to encourage rigorous, transparent, and repeatable research. This paper presents a structured analysis of the track from 2020 to 2025, covering 51 accepted papers. We classify contributions by type and analyze common patterns in datasets, algorithms, frameworks, and evaluation practices, with the goal of understanding how reproducibility is operationalized in practice within the community. Our findings reveal three main trends. First, the track has expanded in scope, evolving from a focus on reproduction and replication to include benchmarking, resources, and methodological contributions. Second, reproducibility papers exhibit a consistent methodological profile, relying
arXiv:2607.27182v1 Announce Type: new Abstract: Question-answer (QA) pairs are widely used in knowledge base construction, question-answering systems, and the post-training of large language models (LLMs). However, important knowledge in long documents is often distributed across multiple paragraphs and connected through complex entity relationships. Such fragmented and relational knowledge poses substantial challenges for existing QA generation methods, which often fail to adequately cover core document content, cross-paragraph semantic connections, and multi-entity relationships. We present GraphQAG, a knowledge graph-guided visual analytics framework for generating high-quality QA pairs from long documents. GraphQAG follows a three-stage workflow. First, it constructs a document knowledge graph by segmenting the document into paragraphs and extracting salient entities and relations. Second, it builds a graph-based generation space from entities, relations, and multi-hop paths to con
arXiv:2607.27125v1 Announce Type: new Abstract: Amateur badminton players increasingly record matches, yet existing tools provide only aggregate statistics or generic summaries, leaving most unable to extract tactical insights without expert guidance. A formative study (N=8) reveals the need for multi-granularity, video-anchored tactical analysis centered on rallies. We derive a taxonomy of performance issues from national-level athletes' annotations and present TactiPlay, an interactive system that instantiates an expert-taxonomy-guided, rally-level, video-anchored review workflow. The system's analytical pipeline organizes match events into taxonomy-grounded feedback, while its interface links structured reports to rally summaries, video evidence, and court visualizations. A within-subjects study (N=16) shows that TactiPlay elicits more frequent, concrete, actionable, and appropriate reflections than a report-and-statistics baseline. These findings show how organizing reviewed match
arXiv:2607.26827v1 Announce Type: new Abstract: Generative AI tools for creative work tend to be designed around the goal of removing friction, on the assumption that smoother iteration and faster output translate into more value for the designer. We argue, however, that this framing leaves out something important about how design ideation works, namely reflection-in-action. The act of accepting, rejecting and reworking candidate ideas is both a path to a final outcome and the process through which designers develop the rationale that allows them to think with their ideas and to communicate them to others. This becomes particularly important in group ideation, where ideas need to be expressed and explained to others to allow the group to extend, reject or combine them further. We suggest that AI in design ideation might be more usefully thought of as a friction agent for reflection rather than as a smoothing agent for output. This reframing opens up a different role for AI in design id
arXiv:2607.26731v1 Announce Type: new Abstract: Current Tools for Thought (TfTs) treat affect as either friction that slows cognitive progress or a signal to optimise it. Drawing on enactive cognitive science, we argue that affect is constitutive of cognition: it reshapes the trajectory of thinking, not just the speed. We identify two core barriers for Affective TfTs: the lack of Shared Attention (caring, directed attention to the user's mode of engagement) and the lack of Affective Reorienting (the capacity to use emotional moments to open new trajectories rather than reinforcing predetermined ones), and propose three design strategies that address both: Chain of Emotion X Chain of Thought, Affective Mirror, and Prompted Reorienting. The strategies are grounded in empirical findings from a study of a touch-aware conversational agent for embodied craft learning, and are oriented as provocations for future design.
arXiv:2607.26659v1 Announce Type: new Abstract: Galvanic vestibular stimulation (GVS) is widely used to modulate self-orientation, balance, and motion perception; the discriminability of frequency-encoded cues further suggests its potential as a standalone modality for embodied feedback. However, synthesizing GVS waveforms congruent with target events or bodily states remains challenging. GVS waveforms combine current direction, intensity, duration, and onset and offset transitions, yet how these parameters jointly shape users' perceptual and associative responses remains underexplored. To address this gap, we contribute a dataset linking GVS waveforms to free-form experience descriptions, as well as a retrieval-guided generative model for synthesizing candidate waveforms from target descriptions. The dataset comprises 100 GVS waveforms and 1,526 valid free-form sensation descriptions collected from 16 participants. Semantic analysis revealed diverse motion- and force-related sensation
arXiv:2607.26526v1 Announce Type: new Abstract: Visual network analysis leverages network visualization authoring techniques to facilitate sensemaking, serendipitous discovery, and hypothesis verification on network data. However, transferring the same paradigm to immersive environments is non-trivial due to insufficient UI affordance for authoring operations. Researchers have studied combining multiple modalities for interactions, but the high learning curve of such input systems limits their adoption by typical data analysts, let alone for network analytics. In this work, we investigate the advantages and limitations of voice as the primary input modality with a research-through-design (RtD) study, in which we design a system that supports voice-based interactions for immersive network visualization facilitated by Large Language Models (LLMs). Through a user study on social network data analysis with participants from social science and computer science backgrounds, we find that voic
arXiv:2607.26517v1 Announce Type: new Abstract: Research on AI-assisted programming has concentrated on the gulf of execution -- how users write successful prompts. We report a candidate phenomenon, an integration bottleneck, that lies in Norman's gulf of evaluation: a repair-relevant contribution reaches the user and fails to become actionable at the point of receipt. Two cases in an eighteen-case corpus of publicly shared AI-assisted-development accounts report this, from a peer and from the system's own output; both fall at evaluation's interpretation stage, and a third, which would fall at comparison, is reached only on an inferential reading and reported as a boundary case. A within-case contrast is consistent with actionability turning on whether the contribution can be restated as an instruction without an intervening judgement. We report this as a candidate warranting dedicated study, not an established regularity; its evidence base is retrospective author self-reports. On the
arXiv:2607.26443v1 Announce Type: new Abstract: Visualization experiments need a set of "good" stimuli that effectively address research questions and hypotheses. Creating, managing, and deploying stimuli are often challenging, as these tasks require tremendous care. Inappropriate stimuli can make the outcome invalid or uninteresting, wasting both researchers' and participants' resources. As the speed of science increases, better support for stimuli-related tasks is essential, yet we lack a closer look at how visualization researchers deal with them. To understand the experiences of visualization experimenters and guide future improvements, we interviewed 19 visualization researchers with diverse backgrounds and experiences. Our findings describe practices and challenges across the life cycle of stimuli, from exploration and selection through shipment, deployment, and analysis. For example, stimuli management and deployment require tedious manual effort, which does not scale for experi
arXiv:2607.26423v1 Announce Type: new Abstract: As autonomous drone deployments scale from individual units to coordinated swarms, the human operator's role shifts from direct piloting to high-level supervision. Current interfaces often treat multi-drone control as a scaled-up version of single-drone operation. We instead investigate how reframing fleet supervision as spatial interaction can better support the spatial, temporal, and safety demands of complex missions. We present FleetScape, a Mixed Reality (MR) sandtable system that externalizes layered real-time mission, safety, and environmental data while enabling fluid transitions between manual intervention and autonomous supervision. We developed a high-fidelity building inspection simulation that generates and streams synchronized multi-drone and environmental data for MR visualizations. We used this prototype to conduct a user study with six experienced drone pilots managing fleets of up to 15 drones. Our findings show that Fle
arXiv:2607.26401v1 Announce Type: new Abstract: Sonomyography (SMG) enables continuous device control via ultrasound-measured muscle deformation signals, but existing SMG interfaces generally require substantial user- and sensor-location-specific training data and provide only one proportional signal or task-specific classification. We present a real-time, sensor-placement-agnostic SMG control system based on sparse optical flow tracking that enables continuous 1-DOF control after minimal calibration (3 pose definitions). We also present a preliminary expansion of this method that augments this algorithm with a short computer-aided calibration to enable 2-DOF control. We evaluate both 1- and 2-DOF systems' performance for a preliminary cohort of 3 cervical spinal cord injury survivors and 6 uninjured individuals across 6 sensor placements spanning the arm, neck, and upper torso. As assessed by a cursor trajectory tracking task, all participants achieved continuous 1-DOF control at all
arXiv:2607.26338v1 Announce Type: new Abstract: Artificial Intelligence (AI) is transforming higher education, but its benefits can vary depending on where, how, and how often it supports learning. While prior research emphasizes cognitive and academic outcomes, this study examines how AI chatbots support the psychological needs and motivational states of engineering students. A survey of college engineering students (n = 206) examined perceived effects of AI chatbots on autonomy, relatedness, and relief from competence frustration. Structural equation modeling with latent interaction effects examined how baseline autonomy, competence frustration, relatedness, and personal agency contributed to perceived AI outcomes. Results indicate that students perceived that AI provided the greatest benefits as relief from competence frustration, smaller benefits for autonomy, and the weakest benefits for relatedness. Baseline motivational states mattered more than demographic factors, and inattent
arXiv:2607.26322v1 Announce Type: new Abstract: People can often understand and use novel artifacts after only a few interactions, suggesting that design choices communicate underlying affordances and causal structure. We propose a formal account of this process by framing cooperative, user-centered design as a cooperative game in which the user is the principal and the designer is an assistant. Inspired by prior work on pragmatic communication (e.g. RSA), our model treats a designer's design decisions as communicative signals and predicts user judgments via recursive mentalizing: designers make design decisions to trade off informativeness about the artifact with efficiency, and users infer the true model of the artifact by inverting this cooperative designer model. We evaluate the model in a design game where designers place visually identical keys on trays to help a user infer which keys unlock which doors in grid-world layouts. We find that pragmatic designer and user models better