Skip to content
AI NEWS JOURNAL
AI, IN PERSPECTIVE

Search AI news

Search headlines, summaries and articles to connect the dots.

Newest first

TileCache: Serving Multi-Round Diffusion Image Editing with Compressed Activation Caches

arXiv:2611.20246v1 Announce Type: new Abstract: Multi-round diffusion image editing can reduce computation by reusing intermediate activations, but retaining these activations across editing rounds introduces substantial memory and bandwidth overheads. Compression can reduce these overheads, but efficient cache construction is challenging: compressing only after editing requires buffering over 100 GB of activations per request. Moreover, the compressed cache must support selective reuse, as future masked edits require only a subset of the cached activations. It also exposes a scheduling challenge, as cached edits are faster than initial generations, causing naive policies to severely delay either initial generations or editing requests. In this paper, we aim to address both challenges through compression-informed serving. We develop TileCache, a serving system for multi-round diffusion editing. It compresses reusable activations and incorporates a scheduler that balances both edits and initial generation requests. Compared with the state-of-the-art image editing baseline, TileCache sustains over 2x the request arrival rate at 90% SLO attainment. It completes edits 2.4-4.5x and initial generations 1.3-2.2x faster than the baseline, while preserving image quality.

Top-Level Operational Equilibrium: When is the response of a distributed system predictable from its own telemetry?

arXiv:2611.20780v1 Announce Type: new Abstract: Distributed systems deployed across the edge-cloud continuum are adapted continuously: scaled, re-routed, given more or less capacity. Each adaptation rests on an expectation of how the system will respond, yet operators cannot tell from the telemetry they already collect whether that response is predictable. We define top-level operational equilibrium (TLOE), a notion of equilibrium for a running system, stated relative to the controls its operator declares. TLOE has three conditions: stationarity with time-scale separation, sufficiency of a top-level aggregate for the controls, and time-reversibility of that aggregate. We prove that an aggregate satisfying the last two exists if and only if the minimal sufficient statistic of the controls is time-reversible, so the coarsest admissible aggregate is derived rather than chosen. Under the first two conditions, the response to a small control change follows from passive telemetry through the steady-state fluctuation-response identity, with a conjugate observable fixed by the type of control: the total work in progress for a change of load, the occupancy of a stage for a change of its capacity. For chains of exponential single-server stages with Poisson input, TLOE holds exactly for load control at every stable throughput, but never for a control on the capacity of the first or last stage. On reference models, a per-stage conjugate overestimates the early response about threefold, blocking between stages breaks both conditions weakly, and a hysteretic autoscaler leaves no admissible aggregate at any reaction time computed. We specify a test for each condition, show that a static sufficiency test with an estimated top-level conjugate has no power, and quantify on simulated telemetry the window lengths the prediction and the reversibility test need. The evidence is on reference models; real telemetry is left to future work.

Two-Robot Evacuation from a Unit Disk in the Segment Visibility Model

arXiv:2611.19992v1 Announce Type: new Abstract: Under segment visibility, two robots discover a target as soon as it belongs to the segment joining their positions, even if neither robot has physically reached it. We study two unit-speed robots starting at the center of a unit disk that must both reach an exit (target) placed at an unknown point of its boundary. In our model robots are permitted to leave the disk while searching. We give a three-phase tangent-and-arc algorithm whose exact worst-case evacuation time is 2\sqrt{3}+\pi/3 \approx 4.511299$. We also prove a matching lower bound for the class of descending horizontal full-slice sweeps, allowing asymmetric robot trajectories. The proof uses an exterior path-length inequality and a feasible minimum-half-width symmetrization. Determining the optimal time over all segment-visibility algorithms remains open.

UMM-JEPA: Generative Pretraining via Masked Next-Scale Representation Prediction for Unified Multimodal Models

arXiv:2611.18668v1 Announce Type: new Abstract: Unified multimodal models can support both visual understanding and generation, yet generative training does not necessarily improve understanding. We argue that this limitation may arise from a mismatch between the representation targeted by generation and the perceptual representation used for understanding. Conventional generation objectives primarily optimize synthesis-oriented spaces such as discrete visual codes or generative latents. We therefore introduce UMM-JEPA, which reformulates generation as progressive prediction of continuous visual-understanding representations. The model constructs a multi-scale representation hierarchy from coarse to fine, while pixel rendering is decoupled from representation pretraining and attached only afterward. Controlled comparisons show that the representation-prediction objective accounts for most of the understanding improvement, while a data-matched control leaves the group means nearly unchanged. UMM-JEPA reaches a 16-benchmark visual mean of 51.71, against 46.36 for the understanding-only Qwen-VL baseline, and the text mean also rises. Compared with discrete-code and latent-flow generation objectives, perceptual representation prediction yields stronger and more consistent transfer to understanding. Linear probes and visual-evidence interventions further reveal richer object and spatial structure and stronger use of image-specific information. These results suggest that generation can benefit understanding when it directly supervises the perceptual representations that understanding consumes.

VidEngram: Agentic Long-Term Video Understanding via Structured Hierarchical Memory

arXiv:2611.18578v1 Announce Type: new Abstract: Long-video question answering is wasteful and lossy when a multimodal model must re-encode the same video for every query: context limits force sparse sampling or truncation, and no reusable representation remains for later questions. We present VidEngram, an agentic video QA system that builds a persistent hierarchical memory once at ingest and answers later questions from that memory. VidEngram captions video and audio jointly with Qwen2.5-Omni, consolidates segment memories into episode and entity layers, and uses a text-only agentic planner with specialized retrieval tools to produce grounded answers. The web interface makes this process inspectable: users can watch memory formation, inspect retrieval traces, verify cited evidence in the video player, and explore entity relationships. On the full OmniVideoBench, VidEngram with a GPT-5-mini planner outperforms a strong single-pass Qwen2.5-Omni baseline (37.0% vs. 32.3%), with gains concentrated on longer videos where single-pass coverage becomes sparse. Because perception is paid once, subsequent queries use low-cost text-only planning.

AI-Voyager: A Public-Interest System for Human-AI Civic Reasoning to Contest and Shape the Rules That Govern AI

arXiv:2611.17564v1 Announce Type: new Abstract: Governments are writing rules for artificial intelligence that decide what counts as an automated decision or a harm, often before the people they affect have a chance to question them. We introduce AI-Voyager, a public-interest system for human-AI civic reasoning, in which people and AI work together to understand, contest and shape the rules that govern AI. A resident can find their city's rules on AI in hiring or upload their country's AI strategy, read it as an affected worker would, open the passage behind any answer, and draft a revision. Its architecture turns 3,653 governance documents from city, state, federal and national governments into passages that keep their page and edition, and retrieves evidence through bounded agentic search that verifies each statement. On 360 new questions that require finding the relevant documents, it reached 83.0% macro recall, against 80.5% for the strongest baseline, 73.7% without iterative search and 99.1% on a benchmark with fixed documents. In a study with 24 participants new to two AI bills, most proposed changes asked how provisions would apply. Participants chose lenses for the kind of explanation they wanted, and citations signaled grounding but rarely led them to check the source. By binding AI assistance to the exact text of public documents, this work introduces a grammar of operations and a design space that inform how such systems are designed and evaluated, and how people take part in governing AI.

An Operationalization of Opaque Serial Depth

arXiv:2611.17495v1 Announce Type: new Abstract: Chain-of-thought (CoT) is currently a valuable tool for overseeing AI models, but some architectural shifts could significantly reduce CoT monitorability. We operationalize a measure that serves as a proxy for the amount of unverbalized serial cognition a model can perform, as a specific instantiation of the notion of opaque serial depth: roughly, the length of the longest path through a model's computation that does not pass through an interpretable bottleneck. Our main contribution is a particular standard for what counts as such a bottleneck. We define natural-language-rooted (NL-rooted) nodes as nodes that were initialized from a pre-training prior to produce human-understandable outputs, whose output space has not been expanded beyond that prior, and whose outputs the model has not been modified to interpret as a different type of data. By construction, the resulting measure, NLS depth, can be adjudicated before training from an architecture, training recipe, and deployment configuration alone, and is insensitive to ordinary post-training such as prompting, distillation, and RL. We find that, across open-weight models, NLS depth correlates strongly with layer count, with some exceptions, and has grown slowly over recent years. It would likely grow slowly under continued scaling of standard transformers, while latent-reasoning architectures such as full-bandwidth transformers could increase it by a large amount. We work through how NLS depth applies to a range of existing and proposed architectures, discuss alternative operationalizations, and give a worked example bounding the depth of Gemma3-1B. We argue that AI developers should track and report the NLS depth of their models as a complement to direct measurements of CoT monitorability.

Availability-Aware Historical Feedback Learning for Electric Arc Furnace Temperature and Chemical Composition Prediction

arXiv:2611.17463v1 Announce Type: new Abstract: Electric arc furnace (EAF) steelmaking plays an increasingly important role in recycled and lower-carbon steel production. However, variability in charged materials, energy input, and operating conditions makes accurate prediction of molten steel temperature and final chemical composition challenging. Intermittent or delayed measurements limit the information available for timely prediction. This study develops an availability-aware historical-feedback learning framework for predicting EAF endpoint temperature and the final concentrations of twelve chemical elements. Eight industrial data sources are integrated into three progressively enriched predictor representations comprising 19 static, 39 temporal, and 47 history-informed features. Historical outcomes are incorporated only when eligible under an assumed 60-minute measurement-release delay. Using 19,738 temperature heats and 3,341 chemistry heats, twelve tuned regression algorithms are evaluated through chronological forward validation. The proposed historical-feedback representation reduces root mean squared error (RMSE) in 147 of 156 matched algorithm-target comparisons. For temperature prediction, ExtraTrees achieves an RMSE of 12.290 degrees C, representing a 5.54 percent reduction compared with its temporal-only counterpart. Historical-feature ablation and feedback-delay sensitivity analyses demonstrate the importance of recent outcome information and its availability. In a separate retrospective later-period evaluation, lightweight adaptive residual correction reduces temperature RMSE from 12.502 to 12.275 degrees C without retraining the underlying model. Exploratory deep learning comparisons and target-specific feature-attribution analyses provide additional insights into predictive behavior. The results establish historical outcome availability and sequential feedback as important considerations for EAF endpoint prediction.

Benchmarks Don't Talk Back: Learning Through Self-Disagreement

arXiv:2611.17563v1 Announce Type: new Abstract: A benchmark usually measures a language model alone with a question, and deployment rarely looks like that: answers get contested, and models give up correct ones under counter-reasoning that is itself wrong. We show that what fails there is not a shortfall of accuracy but a capability in its own right -- knowing when to hold a position and when to yield to a better one -- which we call dialectical competence. It stands on first-answer accuracy without following from it: a model can be made materially better at answering and no better at this, and training for accuracy alone can drift against it. Nor can it be bought at inference, by self-reflection or by sampling more. Taking a cue from the dialectical tradition in philosophy, in which productive opposition arises within a position rather than being imposed on it from outside, we hypothesize that this capability can be trained from opposition the model supplies itself. It states a position, generates the strongest case against that position -- its antithesis -- rules on whether the case stands, and is trained on that ruling, with no second model at training or at deployment. We call this learning through Self-Disagreement: reinforcement learning from verifiable rewards throughout, where the same check that scores an answer against its reference scores the ruling too, so no reward model, no preference data and no human annotation enter anywhere. The capability is trained rather than elicited: weak in the base model, it arrives with the reward, is not paid for out of first-answer accuracy, and holds against challengers the model never trained against, including ones stronger than itself. The opposition a model needs in order to improve, it turns out, need not come from outside it.

CamSpokenNLI: A Pilot Resource for Cross-Variety Transfer Evaluation in Cameroon Pidgin English and Camfranglais

arXiv:2611.17573v1 Announce Type: new Abstract: Cameroon Pidgin English (a West African Pidgin English variety) and Camfranglais are widely used in Cameroon, yet to our knowledge, neither has a dedicated spoken Natural Language Inference (NLI) resource. We introduce CamSpokenNLI, a pilot dataset of 114 scenario-matched spoken NLI items in each variety (228 in total), recorded by one native speaker from Buea. Each item pairs premise audio and text with a hypothesis, an NLI label, and English glosses. The author assigned labels, so the dataset is an exploratory resource rather than a mature benchmark. The NLI experiments use the reference text; the audio is evaluated separately. XLM-RoBERTa baselines with five-fold item-level cross-validation reach 44.7% on Pidgin and 46.2% on Camfranglais (chance 33.3%). Still, neither is statistically distinguishable from hypothesis-only input, so hypothesis wording may account for much of the accuracy. Under disjoint-item cross-variety transfer, where a test item's counterpart is never seen in training, accuracy is 40.1% (Pidgin to Camfranglais) and 45.6% (Camfranglais to Pidgin). A counterpart-leak audit raises this to 60-64%, illustrating how counterpart leakage can overstate transfer. Zero-shot Whisper transcription of the recordings performs poorly: the best word error rates are 66.6% for Pidgin and 45.5% for Camfranglais, and with automatic language identification, Whisper Large-v3 labels 93 of 114 Pidgin recordings as Yoruba or Swahili. Together, the results suggest that such resources need disjoint-item evaluation and hypothesis-only controls, and that speech recognition must be adapted to these varieties before spoken NLI can be tested end to end.

Can You Trust Your Finance Agent? Drift, Diagnosis and Auditable Intervention on One Shipped Financial Research Agent

arXiv:2611.17939v1 Announce Type: new Abstract: Can you trust a finance agent you did not build? We take the reference agent shipped with the Vals Finance Agent Benchmark v2, leave it untuned, and ask four questions a leaderboard score does not answer. Is the score stable? Seven identical runs on 18 questions pass between 105 and 123 of 149 rubric criteria, and the bill varies by up to 17%. When the agent fails, who owns the fix? We classify 20 failed trajectories, from three runs of nine held-out questions, by the layer whose owner would remedy them, under a taxonomy frozen in advance: 13 belong to the agent's instructions, 3 to its tools, 4 to the model, none to its data. Can that diagnosis be trusted? Four blind readers from three model families, given only the traces, name the same fault as our LLM analyzer in 94% of cases, the rate at which they agree with one another (93%). Can a change be trusted? An optimization loop reads the traces, makes one targeted change per round and keeps it only if it clears a fixed threshold. Under a cost objective its final configuration costs 0.34x the baseline on the held-out questions (95% CI [0.22, 0.58]) at accuracy within noise. Under an accuracy objective it gains 14 points on the development questions and 1.5 on the held-out ones (CI [-5.3, +7.2]). Judged against the measured run-to-run noise, under a rule written before the test was opened, the first is a real saving and the second is not. This is a case study of one agent and nine test questions, not a general result.

CarryOn: Learning What to Keep in Context for Long-Horizon LLM Agents

arXiv:2611.17808v1 Announce Type: new Abstract: Long-horizon LLM agents can coordinate tools and applications to complete complex, multi-step tasks. As their interaction histories grow, effective context management becomes essential for reliable and efficient execution. Existing methods reduce context through summarization, pruning, or history selection. However, shorter contexts do not necessarily support better decisions: summaries can omit exact details needed for later actions, while retaining entire interactions can carry substantial irrelevant information. We propose CarryOn, a lightweight context manager that pairs compact execution state with fine-grained, source-linked evidence. A learned selector and adaptive retention policy keep relevant evidence available without additional LLM calls for context management. We introduce two complementary evaluation metrics: $S_\alpha$ prioritizes task success while accounting for resource use, and normalized cost per success (NCS) measures the resources spent per successful task. Across AppWorld, OfficeBench, and Tau2 Telecom with three LLM Actors, CarryOn achieves 75.7% macro-averaged task success, exceeding the strongest baseline by 6.7 percentage points. It also achieves the best average scores on both metrics.

Creating Steering Vectors via Input Optimisation

arXiv:2611.17481v1 Announce Type: new Abstract: Activation steering influences a language model's behaviour by adding a steering vector to its residual stream during generation, but it requires white-box access to modify the hidden states at every generation step. We introduce ActGCG, a modification of Greedy Coordinate Gradient which optimises a suffix to match the target activations obtained by adding a steering vector to the residual stream. We further introduce universal suffixes trained across many prompts at once, a model-translation step that rewrites the suffix into human-readable text midway through optimisation, and a soft-suffix variant to remove the discrete token space constraint. We study two behaviours: steering a model to answer in German, and anti-refusal steering for jailbreaking. For German, a discrete suffix reproduces the steering vector's behaviour almost perfectly and transfers well across prompts and models. For anti-refusal, the equivalence is weaker, succeeding under a phrase-avoidance metric but only partially under the stricter HarmBench classifier. Our results show that a steering vector's behaviour is reachable from the input, with no access to model internals at generation time, up to a loss that depends on the concept being steered.

Creative Product Design under Brand Image Constraints

arXiv:2611.17762v1 Announce Type: new Abstract: Industrial design requires creative visual exploration while satisfying constraints that are often product-specific and implicit. Existing text-to-image (T2I) approaches can generate creative designs through concept blending or multi-agent self-enhancement, but they provide limited mechanisms for ensuring that generated designs remain consistent with such constraints. We focus on brand image as a representative design constraint and propose a framework for brand-aware redesign generation. Given a product name, the framework constructs a brand image survey report that summarizes product and market information, derives a creative direction from the report, and integrates this direction into a self-enhancement loop in which generated images are critiqued and prompts are revised. We evaluate the framework on 50 redesign tasks across five product categories, comparing it with Naive T2I, IP-Adapter/IT-Blender, and unconstrained self-enhancement under the same T2I backbone. By a VLM-as-Judge protocol to measure creativity and brand identifiability, the proposed framework achieves a more favorable trade-off between the two objectives, with particularly clear gains in highly constrained categories such as automotive and motorcycle, where the plausible design space is relatively limited compared with that of other product categories. These results suggest that dynamically eliciting and textualizing brand image can guide creative product design generation under brand image constraints.

Does the Ranking Hold? A Practitioner-Facing Audit of Hallucination Detectors on Small, Quantized, On-Device LLMs

arXiv:2611.17339v1 Announce Type: new Abstract: Hallucination detection for large language models is dominated by methods evaluated on 7B-70B research models with abundant sampling budgets and exact-match or ROUGE correctness labels. The small models practitioners actually ship on-device, Phi-3-mini, Llama-3.2-3B-Instruct, Qwen2.5-3B-Instruct, and Gemma-2-2B-Instruct, typically at 4-bit precision, sit outside this evidence base. We present a reproducible one-GPU audit of four detectors spanning three paradigms: length-normalized token-margin calibration, a lexical n-gram baseline, and two sampling-based consistency methods (semantic entropy and SelfCheckGPT-NLI) on two small open instruction-tuned models (Llama-3.2-3B-Instruct and Qwen2.5-3B-Instruct) loaded at 4-bit NF4 quantization on a single Colab T4 GPU, evaluated on closed-book TriviaQA. Cross-model ranking transfer is poor: Spearman rho is -0.80 and Kendall tau-b is -0.667, meaning the detector ranking actively inverts between the two models, and the cheap lexical baseline is statistically indistinguishable from the sampling-based consistency detectors on both models under bootstrap 95 percent confidence intervals. As a second, methodological contribution, we show that the standard token-overlap-to-multiple-choice correctness labeler for TruthfulQA-Generation produces near-chance AUROC (0.46-0.59) for all four detectors on both models, a labeler failure, not a detector failure, since the same detectors reach 0.69-0.89 AUROC on TriviaQA under the identical pipeline. We release the harness, prompts, and per-item scores and frame the detector choice explicitly as a fairness-of-error-rates allocation problem for practitioners deploying on resource-constrained, on-device hardware.

Fair and Efficient Distributed EV Charging Coordination Using Threshold Cryptography and Machine Learning

arXiv:2611.17577v1 Announce Type: new Abstract: Distributed coordination of electric vehicle (EV) charging with the alternating direction method of multipliers (ADMM) keeps EV constraints private but assumes that every participant follows the algorithm. This assumption fails in practice, where a biased operator or a man-in-the-middle (MITM) attacker can send inconsistent signals to different EVs, and compromised EVs can deviate from their prescribed updates. We first prove that attacks by such arbitrarily malicious, i.e., Byzantine, participants do not prevent convergence and therefore remain undetected, and derive the most damaging attack in closed form. We then propose a framework that combines a threshold-signed broadcast with upload recomputation, which detects every deviation without false alarms at linear instead of quadratic communication cost. To keep the optimization running, each affected EV recovers locally with a certified, fleet-aware machine learning price forecast, and we prove that the optimization remains convergent with a recovery cost bounded by the shape of the forecast error. Simulation results with real-world load, price, and charging data show that a biased operator raises the annual cost of disadvantaged EVs by 132.7% and lowers that of advantaged EVs by 49.3% without changing the aggregate load, whereas the defense limits all cost changes to 2.6%. An external MITM attacker raises the cost of its victims by 57.6%, the peak load by 5.8%, and the grid cost by 15.8%, whereas the defense keeps them within 1.1%, 0.2%, and 0.3% of benign operation. With up to 70% of the EVs attacked, the recovery converges in at least 97.8% of the runs.

HarmonyText: Harmonious Scene Text Editing via Reverse-to-Refined Pair Synthesis and Region-Decoupled Frequency Learning

arXiv:2611.17733v1 Announce Type: new Abstract: Natural language-guided, mask-free scene text editing aims to modify text in real-world images using only textual instructions, yet remains challenged by unreliable pseudo-GT supervision, over-smoothed text generation, and the absence of GT-free evaluation. Existing methods often rely on imperfect pseudo-targets with background discrepancies, while standard MSE-based velocity regression tends to suppress high-frequency details, leading to blurred text boundaries and weakened local textures. To address these issues, we propose HarmonyText, a mask-free scene text editing framework with region- and frequency-aware supervision. HarmonyText introduces Region-Decoupling Optimization (RDO), which aligns edited text regions with pseudo-targets while anchoring preserved regions to the source image, and Cascaded High-Frequency Texture Guidance (CHFTG), which applies cascaded wavelet-domain constraints to preserve sharp boundaries and fine-grained textures of textual content. We further propose Reverse-to-Refined Pair Synthesis (RePair) to construct STE25K, a high-quality 25K paired dataset, and introduce a GT-free evaluation protocol with Color Distribution Divergence and a Harmony Assessment Model. Extensive experiments show that HarmonyText outperforms existing methods in text editing accuracy, background fidelity, color consistency, and perceptual harmony.

Improved Regret via Variance-Aware Policy Optimization in Adversarial MDPs

arXiv:2611.17469v1 Announce Type: new Abstract: We study episodic adversarial MDPs with unknown dynamics and full-information rewards and achieve regret with leading term $\widetilde O(\sqrt{H^3SAK})$. Here $H$ is the horizon, $S$ the total number of states across decision layers, $A$ the number of actions, and $K$ the number of episodes. This improves the two incomparable prior leading terms by factors $\sqrt{S/H}$ and $\sqrt{H^3}$, leaving a factor $\sqrt H$ to the lower-bound scale. We give a reduction from unknown to known dynamics for \emph{downstream-dependent} algorithms, whose policy at step $h$ is computed only from information at steps $h+1$ and onward. This reduces the remaining leading-order gap to finding a downstream-dependent known-dynamics algorithm with $\widetilde O(H\sqrt K)$ regret, which would yield the tight $\widetilde O(H\sqrt{SAK})$ leading term for unknown dynamics. We instantiate the reduction with VOPO, a policy optimization (PO) algorithm. Its improved guarantees follow from a variance-aware analysis enabled by a novel bonus, which we call the Stability-Matching Bonus (SMB). VOPO achieves $\widetilde O(\sqrt{H^3K})$ regret under known dynamics with full information; its $\sqrt H$ gap to the target $H\sqrt K$ rate is inherited by our unknown-dynamics guarantee. We also show that with known dynamics and trajectory bandit feedback, VOPO achieves $\widetilde O(\sqrt{H^2SAK})$ regret. Both known-dynamics guarantees improve the corresponding state-of-the-art PO bounds by a factor $\sqrt H$. A matching lower bound establishes that the full-information rate is tight in one parameter regime.

In the Eye of the Beholder: Human Label Variation in Textual and Visual Entailment

arXiv:2611.17954v1 Announce Type: new Abstract: Human Label Variation (HLV) is increasingly recognised as an informative signal that reflects meaningful differences in how people interpret the same input. While HLV has been studied extensively in Natural Language Inference (NLI), much less is known about HLV in Visual Entailment (VE; Xie et al., 2019), where the textual premise is replaced by an image, keeping the hypothesis and label space fixed. We introduce VariNLI-VE, a dataset for studying how HLV changes between paired NLI and VE instances. We re-annotate a subset of originally neutral SNLI (Bowman et al., 2015) instances and their visual counterparts, with the same annotators evaluating both the textual and visual versions. Each annotation includes a label, confidence score, and explanation, yielding 2,250 annotations. Labels and explanations are collected jointly, providing ecologically valid explanations. Because the image may reveal information not stated in the text, one might expect disagreement to decrease in VE. Instead, we find no clear overall reduction: disagreement decreases for some instances while it increases for others. Explanations reveal further variation, as annotators can retain the same label while changing their rationale. When evaluating vision-language models (VLMs) on the paired data, we find lower accuracy, weaker alignment with human label distributions, and generally higher uncertainty in VE than in NLI.

LLM-Guided Cluster Lifecycle Management: Deciding When and How to Evolve Behavioral Clusters

arXiv:2611.17616v1 Announce Type: new Abstract: Behavioral clusters learned over user or item interactions drift as behavior evolves: some fragment into sub-behaviors, some converge, some disappear, and new patterns emerge. Two paradigms dominate the maintenance of evolving clusterings. Threshold triggers fire a pre-specified action when a hardcoded statistic crosses a cutoff; refit-and-track (MONIC-style) re-clusters every window and infers transitions from how the new clusters align with the previous window's. Both hardwire the decision into a threshold or a re-clustering rule. We study an alternative in which a large language model (LLM) makes the lifecycle decision in place. Given a threshold-free, whole-population view of the current cluster set (statistical health, labels, member texts, residual mass), the LLM decides per cluster whether and how to act (split, merge, retire, create, or no-action) without re-clustering. Because the useful signal channel depends on how informative the labels are, we fix a richness-conditioned policy in advance: statistics when label distinctiveness is low, fused statistics+semantics when it is high. On three public datasets spanning label richness (YooChoose 0.17, MOOC 0.50, News-Category 0.70), the policy-selected LLM mode obtains the highest F1 on all three, with far fewer interventions and stronger restraint than tuned threshold heuristics and a MONIC steelman (paired Wilcoxon p <= 0.003, bootstrap 95% CIs excluding zero). The gain is concentrated more in precision than recall. The useful modality shifts from statistics to semantics/fusion as richness rises. Semantics hurt on YooChoose, fusion is roughly neutral on MOOC, and the semantic channel is required on News. One base prompt template and one fixed policy serve all three datasets, whereas the tuned baseline collapses when moved between them. We treat this shift as a practitioner diagnostic observed on three datasets, not a validated law.