Papers

Dense linear maps carry much of the parameter and computational burden of modern neural networks, yet their dense form leaves the organization of learned couplings implicit. Quantum many-body physics organizes exponentially large operators by writing a global Hamiltonian as a sum of local terms, \(\hat H=\sum_k\hat h_k\). Whether the same structural principle can carry learned neural maps is unknown. We introduce Tensor Mixture (MixT), which represents a dense map as a natively executable sum of overlapping local tensor operators without imposing an explicit matrix-rank constraint. The local-term count \(N_T\) sets the effective nonlocality and operator complexity, while the number of replaced Transformer blocks \(N_B\) extends this structural coordinate across network depth. Tests on Qwen3-8B and LLaMA2-7B reveal a broad recoverable regime followed by an abrupt, model-specific boundary that is remarkably stable against changes in \(N_T\). Accuracy and output-distribution statistics reorganize together across the boundary; in LLaMA2-7B, the same depth separates two scaling regimes of inter-layer geometry drift. The directly executed structure also reduces parameters, arithmetic, storage, and memory. These results establish the local-sum structure as a viable organizing principle for learned linear maps at billion-parameter scale and expose a sharp boundary in their tolerance to structural simplification.

Aligning large language models (LLMs) with diverse and multifaceted user preferences is a fundamental challenge in personalized AI systems. Existing multi-objective alignment methods either rely on costly training or require pre-trained reward models for each preference, making it difficult for them to adapt to evolving preferences. Prompt-based personalization offers a training-free alternative, but prompting alone often provides limited steerability, as LLMs may overemphasize or overlook certain preferences and fail to give users reliable control over the relative importance of different objectives when conflicts arise, leading to suboptimal alignment. In this paper, we introduce MATO, a training-free framework for Multi-objective personalized Alignment with Test-time Optimization. MATO formulates personalization as a test-time optimization problem that steers the relative importance of multiple objectives through controllable weights during decoding, without modifying model parameters or requiring external reward models. Specifically, a reward discovery module recovers preference rewards directly from the backbone LLM for diverse objectives specified in natural language, while a weight optimization module dynamically adjusts objective weights based on the user's initial preferences and the partially generated response to balance competing objectives during generation. The resulting rewards and weights jointly guide an online optimization procedure over the token distribution, enabling better alignment with the target objectives. Extensive experiments across multiple datasets and backbone LLMs show that MATO consistently outperforms strong baselines, achieving Pareto-improving multi-objective alignment and stronger steerability. These results highlight test-time optimization as a promising direction for scalable, controllable, and model-agnostic personalized alignment.

We introduce P1SCO, a dataset of social media comments collected from three distinct platforms, annotated according to ten social dimensions to capture the diversity of social interactions and perceptions. The dataset is carefully disaggregated to allow analysis at the level of individual comments, annotators, and platforms. In addition to the social dimension labels, we include rich metadata on the annotators, including demographics, Big Five personality profiles, and political affiliation. This combination of comment-level annotations and annotator-level features enables nuanced analyses of how social perception varies across platforms, individual differences, and demographic factors. By preserving the diversity of annotator perspectives, our dataset supports studies of inter- and intra-annotator agreement, the influence of personality and political orientation on social interpretation, and the cross-platform dynamics of social discourse.

Tool-using LLM agents produce trajectories whose calls form a directed dependency graph: earlier tool outputs supply arguments to later calls. Whether this execution structure is represented inside the model is unknown; prior structural probes have targeted static code or chain-of-thought text, not an agent's run-time call graph. A low-capacity edge probe on the residual stream of Qwen3-32B decodes the tool-call dependency graph well above both a Hewitt–Liang random-label control and a positional baseline. A counterfactual contrast between value corruption and structural perturbation indicates the signal tracks abstract topology rather than identifier values, and replicates under an independent, non-substring oracle. The non-positional component replicates on three further interactive multi-hop benchmarks and attenuates as call order alone becomes a sufficient proxy for dependency, vanishing in single-shot planning. Per-layer activation patching shifts the probe at a later, non-patched boundary, evidence that the representation propagates rather than passively reads out, though the realised tool call does not move. To our knowledge this is the first structural probe of an LLM agent's runtime tool-call dependency graph. Our claims concern representation, not behavioural control, and span two model families and one primary domain.

Evaluating mobile user interface quality at scale remains a persistent challenge in software engineering and human-computer interaction. Rule-based heuristic methods offer structural reliability but demand significant engineering effort, while human annotation does not scale to the volume of applications produced annually. Large language models present a promising alternative, yet their reliability for structured UI judgment has not been systematically examined, and the patterns behind their failures remain insufficiently characterized. This paper addresses both gaps. We begin with the complete RICO dataset of 66,261 real-world mobile application screens, from which we derive a refined evaluation corpus of 15,000 screens through a rigorous, literature-guided selection process. Each screen is assessed across seven criteria: structural JSON validity, minimum visible element count, clickable component presence, non-zero layout bounds, image integrity, and perceptual duplicate removal. Against this corpus, we apply a heuristic baseline built from severity-weighted usability signals, normalized layout metrics, and pixel-ratio complexity measures calibrated to real user sentiment. Multiple language models independently rate each screen across usability, layout quality, and visual complexity from structured JSON descriptions and raw screenshots. Dimension-level comparison against the heuristic uses agreement rates, Cohen's Kappa, and confidence calibration. Recurring divergence patterns are organized into a failure taxonomy and interpreted through transformer architectural signatures: MLE plausibility bias, attention misgrounding, and autoregressive over-commitment.

In biological evolution, unconstrained mutation can lead to catastrophic outcomes: organisms may evolve enhanced capabilities while losing essential functions for survival. Nature's solution is developmental constraints, where core regulatory genes remain anchored while peripheral genes adapt freely. We observe that current self-evolution algorithms for large language models lack analogous constraints. They optimize purely for capability, implicitly assuming safety will be preserved. Our experiments reveal this assumption to be dangerously wrong: models can misevolve into powerful yet dangerous entities. Inspired by how Hox genes anchor body structure across $500$ million years of evolution, we propose Circuit-Anchored Evolution (CAE). Using mechanistic interpretability, we identify a tiny safety circuit, comprising less than $2$% of model features, that causally mediates safety behaviors. We anchor this circuit during evolution, constraining it within a small displacement bound while allowing the remaining features to evolve freely. This mirrors the biological principle of evolvability with constraint: preserving what is essential while adapting what is peripheral. Experiments across $3$ model families and two evolution algorithms demonstrate that CAE achieves superior safety preservation with minimal capability loss, substantially outperforming explicit reward-based constraints in both effectiveness and efficiency. Just as developmental constraints prevent biological evolution from producing nonviable organisms, circuit anchoring prevents model evolution from producing capable but dangerous systems.

Sampling in LLM inference comprises a combinatorial set of logit processing, token selection, and verification operations for speculative decoding. However, existing implementations either accelerate only subsets of this pipeline, rely on multiple kernel launches, or assume homogeneous sampling behavior across a batch, limiting support for dynamic serving workloads and preventing efficient CUDA Graph execution. We present $\textbf{SonicSampler}$, a unified suite of tile-aware Triton kernels that vertically fuses the complete sampling pipeline into a fixed, workload-aware execution model. Our kernels support dynamic per-request sampling behaviors, including grammar-constrained decoding, repetition, frequency and presence penalties, logit bias, temperature scaling, top-$k$ / top-$p$ / min-$p$ filtering, and speculative verification - within a single batched kernel while remaining fully CUDA Graph-compatible. Central to our approach is a novel hierarchical two-stage top-$k$ algorithm that achieves up to $\textbf{10x speedup}$ over competitive baselines and exploits the low-entropy structure of LLM outputs to enable efficient selection over large vocabularies. Across heterogeneous speculative decoding workloads, SonicSampler achieves up to $\textbf{16x speedup}$ over state-of-the-art baselines while preserving flexible batched execution.

Joint Embedding Predictive Architectures (JEPAs) have reshaped self-supervised representation learning in vision. The recent LLM-JEPA ported JEPA to autoregressive language models but inherited two steep costs from the causal-attention substrate: it demands explicit multi-view data (e.g., text-code pairs), and it requires two gradient-carrying forward passes per step. We introduce DLLM-JEPA, which pairs JEPA with masked-diffusion language models to eliminate both costs at once. The bidirectional attention of diffusion models yields two semantically distinct views of the same input via different masking rates – no explicit pairs needed – and supports a single gradient-carrying forward pass, cutting training FLOPs by 33% relative to LLM-JEPA. DLLM-JEPA improves over diffusion-only fine-tuning in every (task, architecture) combination we evaluate: up to +18.7 pp on LLaDA-8B GSM8K and +11.4 pp on Dream-7B GSM8K, with consistent positive gains on Spider, NL-RX-SYNTH, and Django. Beyond accuracy, DLLM-JEPA exhibits a dual-win property: on LLaDA-8B with the Wide-t configuration, it simultaneously raises GSM8K accuracy (67.1 vs. 65.2, +1.8 pp), drives held-out Wikitext loss below the pre-trained base, and preserves MMLU accuracy at base level across three fine-tuning seeds – whereas an L2-to-base parameter anchor matches baseline accuracy with no task gain. Layer-wise probing reveals the mechanism: a geometric-functional drift dissociation in which the fine-tuned backbone moves further from the pre-trained weights than the baseline yet forgets less on held-out Wikitext, with the amplification concentrated in middle transformer layers. The pattern appears on Dream-7B as well, indicating the phenomenon is not specific to a single backbone.

Agentic RAG systems that equip language models with dozens to hundreds of tool definitions face a critical resource conflict: tool schemas consume the same context window needed for retrieval-augmented generation. We present the first systematic study of this tool-context trade-off, evaluating 14 models spanning 1.5B-32B local models plus one frontier API model across 6,566 controlled API calls at three context budgets (8K, 16K, 32K) with 28 tool definitions. Applying TSCG conservative-profile compression (44-50% schema token savings), we observe a binary enablement effect: at 8K tokens, JSON-schema tool definitions overflow the context window entirely, yielding near-zero EM (2.6% average), while compressed schemas restore RAG functionality with +20.5 pp average exact-match lift across all eight models (+24.7 pp among the six exhibiting full enablement). At 32K – where both formats fit – four of five tested models show delta <= 1 pp, confirming the effect is purely budget-driven. External validation on HotpotQA (50 multi-hop questions) shows +48 pp EM under the same overflow scenario. Frontier scaling tests demonstrate that JSON schemas overflow at ~494 tools while compressed schemas remain operational beyond 800 tools. Our results establish tool-schema compression as a necessary infrastructure layer for agentic RAG in constrained-context deployments. All code, data, and checkpoints are publicly available.

Effective features are crucial for predictive model performance, but creating them often requires domain expertise, limiting scalability across applications. We define feature engineering as an agentic code generation problem: features are not static data transformations, but executable programs that can be generated, evaluated, and iteratively improved. We present Eureka, an LLM-driven framework with three stages. (1) An Expert Agent, fine-tuned via SFT on domain knowledge, produces structured feature design plans in JSON format. (2) An LLM Feature Factory translates each plan into executable Python code through chain-of-thought reasoning, turning feature hypotheses into runnable programs. (3) A Self-Evolving Alignment Engine uses Reinforcement Learning (GRPO) with dual-channel reward (metric-based utility + semantic alignment) to enhance code quality. By expressing features as programs, the learned generation patterns can transfer across domains. Evaluated on 7 public benchmarks in healthcare, finance, and social domains, Eureka consistently outperforms both traditional AutoFE and LLM-based baselines. We further demonstrate Eureka's effectiveness on cloud GPU resource demand prediction at Alibaba Cloud, where Eureka improves demand fulfillment rate by 16% and lowers computing resource migration rates by 33%.

User queries are often underspecified and may admit multiple valid interpretations. Rather than silently making assumptions about the user's intent, a helpful assistant should surface such ambiguity by asking a clarifying question. Doing so requires two abilities: recognizing that a query is ambiguous, and acting on that recognition by seeking clarification instead of answering directly. To study these abilities, we evaluate models on ambiguous, unambiguous, and disambiguated questions in three settings: standard question answering, explicit ambiguity judgment, and behavioral analysis, where a judge model classifies responses as direct answers, refusals, or clarifying questions. We find a clear gap between recognition and behavior: models often identify ambiguity when explicitly asked to judge it, yet in the QA setting they overwhelmingly default to direct answers. Retrieved context further widens this gap by improving answerability while making models even less likely to ask clarifying questions.

Recent advances in large language models (LLMs) have made it increasingly difficult to distinguish human-written text from AI-generated content. Many existing detectors train supervised neural classifiers that achieve strong in-distribution performance but are often opaque and can degrade substantially under distribution shift. We present READER, a reasoning-enhanced AI text detector that outputs both a human/AI label and a structured rationale describing the evidence for its decision. A key component of our approach is READ, a curated supervision set of rationales and verdicts. We fine-tune an LLM on READ to build READER, which reasons before detecting at inference time. Despite having only 1.5B parameters, READER consistently outperforms existing detectors as well as prompted, high-capacity LLM baselines (GPT-5.2, Gemini-3-Pro, and DeepSeek-V3.2), which are 100 to 1000 times larger in scale.

Current language modeling approaches are built around tokens. Text corpora are split into tokens, and models are trained by performing computations on these tokens, such as predicting the next token given the preceding ones as context. This paradigm has become the standard in modern language modeling, especially given the outstanding performance obtained by token-based architectures. However, recent works have not only begun to question how language models process and understand meaning from tokens, but also to question whether using higher levels of granularity could advance the research field. This led to the idea of Concept Modeling, that is, to directly train models for next-concept prediction rather than next-token prediction. In this work, we introduce Mimir, a 1.6B Large Concept Model trained for multilingual concept understanding and generation. We leverage a large-scale multilingual pre-training corpus (38,883,987,240 sentences) spanning 46 languages and a large-scale multi-turn, and multilingual instruction-tuning dataset (66,816,428 sentences) covering a total of 35 languages. We extensively evaluate model performance against a language model with a comparable number of parameters.

Inference time optimization techniques, such as repeated sampling, have significantly advanced the reasoning capabilities of Large Language Models (LLMs). However, the critical role of model uncertainty remains largely underexplored in these optimization strategies. In this paper, we investigate the dynamics of confidence along reasoning trajectories and for first time reveal a surprising and unique pattern: correct answer traces tend to exhibit confidence improvement over time (positive confidence gain), while incorrect traces show attenuated or declining confidence as reasoning proceeds. Based on this observation, we propose Confidence Dynamic Gain (CDG) based voting, which incorporates how the confidence trajectory of the response evolves along the reasoning chain. Experiments across four open-source architectures (DeepSeek-R1, gpt-oss, Gemma-3, Qwen-QwQ) on the AIME24/25, HMMT25, and BRUMO25 benchmarks demonstrate that CDG yields a significant performance boost over baselines. These results demonstrate that our method provides a robust discriminative signal for improving answer selection in LLM reasoning. We also provide theoretical insights for this phenomenon. Code will be released at https://github.com/Accenture/CDG.git.

Two methodologies dominate current practices of benchmarking: rubric-based scoring evaluates items against predefined criteria, whereas comparative judgment elicits pairwise preferences between outputs. Although both methodologies are widely used, the choice between them is rarely justified. We release JudgmentBench, a benchmark of 30 real-world legal tasks, paired with 1,539 rubric scores and 1,530 pairwise preference judgments collected from practicing attorneys–including at major U.S. law firms–with substantial experience. The annotations constitute the first publicly available dataset in a high-expertise domain in which both supervision signals are elicited from the same experts on the same items. Using LLM-generated outputs at three constructed quality levels, we provide an initial empirical comparison: comparative judgments recover the intended quality ordering substantially better than rubrics under both a per-task rank-correlation metric (mean Spearman's rank correlation of 0.908 vs. 0.150, estimated difference = 0.758 [0.494, 1.021]) and a per-judgment pairwise win-rate metric (0.669 vs. 0.542, estimated difference = 0.127 [0.067, 0.186]), while requiring less than half the annotation time. The patterns hold for human annotators and LLM autograders. Beyond this initial comparison, the paired structure of the dataset supports a broader research agenda on how expert judgment should be elicited, aggregated, and used as supervision in domains without verifiable ground truth.

Large language models are widely deployed in high-stakes NLP tasks, yet risks such as bias, hallucination, adversarial vulnerability and unreliable generalization remain. Probe-based auditing reveals inconsistencies in model behavior. Adversarial text generation uncovers robustness gaps, especially in lower-resourced languages with limited benchmarks. Enterprise text-to-SQL settings expose the difficulty of validating outputs over private and large-scale databases. Human supervision is essential for probe validation, adversarial verification and domain-specific annotation, but it is costly and hard to scale. This survey examines recent human-in-the-loop methods that shift NLP from automation toward collaboration for safety and trustworthiness. We review how human expertise supports auditing, robustness evaluation, data construction and model steering. Our findings highlight gaps in scalable probing, sustainable robustness benchmarks, low-resource settings and governance of private systems. We outline practical research directions for adaptive auditing, collaborative evaluation and accountable deployment.

Emotion-Cause Pair Extraction (ECPE) was introduced to explain why an emotion occurs, but this goal is now often reduced to binary pair/non-pair prediction. This proxy is useful for direct-cause extraction, yet easy to over-read as evidence grounded emotion explanation. We show that this interpretation is only partially valid. In IEMO-MECP, 90.9% of original positives remain emo-cause and 95.0% of original negatives remain non-pair, confirming that the binary ECPE task is largely preserved. The problem is that direct triggers alone do not constitute a grounded explanation. Emo-context, an utterance that helps interpret a target emotion without directly causing it, appears on both sides of the original boundary and is enriched near binary uncertainty, showing that the binary boundary has no stable place for such discourse evidence. Across evaluated ECPE models, direct triggers are recovered more reliably than contextual support. Under shortcut pressure, this imbalance becomes consequential. Binary-trained models assign higher pair scores to nearby lexically similar non-pair candidates than to evidence supported but structurally harder emo-cause and emo-context pairs. Thus, pair scores can reward convenient attributions over grounded explanations. High binary ECPE performance indicates that a model can identify direct triggers; it does not indicate that the model has explained the emotion. Code is publicly available at https://github.com/panzhzh/ECPExsame.

Pluralistic alignment requires systems to adapt to diverse user values, communication styles, and contextual assumptions. We believe that a foundational prerequisite for such alignment enabling accurate preference elicitation from people when their intent is under-specified or ambiguous. We study the problem of preference elicitation in multi-turn question answering by decomposing the problem into two components: a clarification policy, which decides whether to ask a clarifying question or answer directly, and post-clarification answering, which produces the correct final answer once the missing information is provided. We show, using the PACIFIC benchmark, that supervised fine-tuning rapidly improves the clarification policy, however, final answer accuracy remains substantially lower even when the model takes the correct action. This gap indicates that understanding and correctly interpreting the user's response is the critical gap in multi-turn question-answering systems.

Travel planning in the real world is overwhelmingly a group activity, yet existing LLM travel-planning benchmarks reduce it to a single user, where the field is approaching saturation. This single-user assumption sidesteps what makes group planning hard for an agent: discovering private preferences across multiple users, surfacing conflicts, and balancing utility against fairness. To bring the task back to its multi-user reality, we introduce \textbf{GroupTravelBench}, the first benchmark for multi-user, multi-turn travel planning. Built from real user profiles, POI data, and ticket prices, it comprises 650 tasks across three difficulty levels, each running in a synchronous group-chat sandbox with cached tool data for reproducible offline evaluation. Beyond the multi-step reasoning and tool use that single-user benchmarks already test, GroupTravelBench probes three group-specific capabilities: (i) elicitation of private preferences through multi-turn dialogue; (ii) coordination of inter-user conflicts via compromise or subgrouping; and (iii) planning that balances group utility against fairness. We pair this with a complementary evaluation framework combining rule-based outcome metrics and LLM-judge process metrics. Across a wide range of frontier models, even the strongest agents fall short on all four rule-based outcome metrics, with plan validity below 12%, suggesting that group-level outcome quality is a key open challenge for LLM travel-planning agents.

Reward hacking arises when a model improves a proxy reward by exploiting shortcuts rather than solving the intended task. We study this failure mode through the geometry of reinforcement learning updates in language models and argue that hacking emerges when optimization drifts away from a stable low-dimensional learning trajectory. We analyze this drift through dominant singular directions of parameter updates and show that reward-hacking runs exhibit substantially larger directional change than clean runs. Motivated by this observation, we introduce trusted-direction projection, which constrains gradients to remain within a clean reference subspace. Across reward-hacking experiments on mathematical reasoning, the proposed approach delays shortcut exploitation and better preserves task performance.

Formalizing legal provisions promises machine-accessible law and automated legal reasoning, and recent LLMs make it tempting to generate such formalizations directly from statutory text. However, any formalization makes implicit interpretive choices whose consequences are hard to anticipate, especially if an LLM is the author. We present a method for systematically comparing different formalizations of the same legal provision by their inferences on individual cases. Given multiple formalizations of a provision, we match them at the node level, derive a shared interface for each pair from the matching, and use a SAT solver to enumerate the edge cases on which any two formalizations disagree. Selected edge cases are then verbalized into concrete factual scenarios that a legal expert can examine and act on. We apply our method to formalizations of ten EU provisions generated by nine frontier LLMs. We find that behavioral divergence between formalizations is essentially uncorrelated with their structural agreement and that the verbalized cases reveal qualitatively distinct types of disagreement, including divergences that mirror genuine controversies in the legal commentary.

Knowledge graph (KG) is an abstraction that can be extracted from text corpora and used for in-depth reasoning. Prior work has leveraged KGs to fine-tune language models (LMs), enabling domain-specific superintelligence. In this work, we explore whether KG-driven in-depth reasoning capabilities can emerge in neuroscience using only information contained within a single authoritative textbook. The central hypothesis is that structured knowledge, when distilled into a high-quality KG and converted into KG-grounded question-answer (QA) supervision, is sufficient to produce expert-level reasoning through a fine-tuned LM that surpasses large language models (LLMs) in accuracy, while employing orders of magnitude fewer parameters. We construct a textbook-derived KG via a dual-LLM validation pipeline, expand it with a masked LM trained on the KG topology, generate multi-hop QA items, which include QA pairs and reasoning traces, to fine-tune an LM exclusively on KG-derived supervision, and apply reinforcement learning using path-derived KG signals as implicit reward models. Our results demonstrate that deep, mechanistic neuroscience understanding can be induced in the model without reliance on large, heterogeneous web-scale corpora. The KG-based synthetic neuroscience curriculum that readers can quiz themselves on, and the fine-tuned LM, are available at the following GitHub location: https://kg-bottom-up-superintelligence.github.io/neuro-bench.

Audio-language models (ALMs) are increasingly used for audio captioning, question answering, and open-ended audio understanding, but their inference cost remains high when audio inputs are represented as long prefix-token sequences. These audio prefixes consume context budget, increase memory usage, and make deployment harder in resource-constrained or latency-sensitive settings. Existing training-free audio-token reduction methods mainly rely on fixed pooling or score-based pruning. Fixed pooling is content-agnostic, while score-based pruning can preserve isolated salient tokens but discard nearby acoustic context. We propose Local Temporal Bipartite Merging (LTBM), a training-free encoder-space compression method that merges similar nearby audio tokens under an explicit temporal window constraint. Beyond introducing LTBM, we use a controlled Global Merge variant to isolate whether temporal locality itself is a useful inductive bias for audio-token compression. Experiments on AudioCaps, Clotho, and MMAU with Qwen2-Audio show evidence of a task-dependent locality effect: locality-aware merging is more favorable for captioning at several compression settings, especially under stronger compression, while global matching is more competitive for multiple-choice audio understanding. A cross-backbone validation on Audio Flamingo 3 further supports the captioning-side advantage of locality-aware merging under moderate and aggressive compression.

In most existing AI humor research, humor was treated as either "present" or "not present." We explore the concept of humor as a social interaction with context and explanations. During this project, we defined a humor reasoning data object and developed a way to prompt LLMs to generate an explanation of humor effective for general population. We iterated from an earlier prompt to an improved prompt, found that the later version reduced important errors, and then scaled generation to a large number of data objects which have the potential to enable data synthesis and data augmentation for AI humor research. Our main takeaway is that better prompting of an LLM improves humor explanation quality, especially by handling missing context, multi-modality, and transcript issues more carefully. These results establish a strong foundation for future work on AI understanding of humor as social behavior. All code and data are available at: https://github.com/anna-arnett/ai-humor/ .

Large language models for vertical domains are bottlenecked by the scarcity of complex, domain-specific task-oriented dialogues. Existing data acquisition pipelines face a persistent trilemma: expert annotation is expensive, real-world service conversations are constrained by privacy and commercial restrictions, and static corpora quickly become temporally stale. We propose Stream, a data-centric framework that leverages publicly available streaming media (live streams and short videos) to synthesize high-value service dialogues at scale. Stream mines authentic interaction signals from noisy streams and synthesizes conversations by integrating role-grounded persona construction with Conversational Blueprint construction; it further adopts retrieval-augmented generation (RAG) to support knowledge-aware responses. Based on Stream, we release StreamDial, a large-scale multi-domain dataset covering Automotive, Restaurant, and Hotel. StreamDial contains 87,498 dialogue sessions and 1,497,320 turns in total, with an average of 17.11 turns per session and a comparable scale across domains. Each session is organized as a structured quadruplet $\langle P_u, P_a, B, H \rangle$ that pairs dialogue history with explicit user/agent personas and a Conversational Blueprint, capturing realistic service behaviors such as requirement mining, constraint conflicts, negotiation, and recovery. Evaluations with automatic judges and downstream tasks show that StreamDial improves intrinsic dialogue quality over strong baselines, and models trained with StreamDial improve Dialogue State Tracking across backbones; we further report a completed human-evaluation set and encouraging multilingual transfer on Qwen3-8B under a controlled training budget. The data is released in https://github.com/hitxueliang/DialogDataSetBySTREAM.

Reliable forecasting of renewable energy generation is a foundational requirement for grid stability energy trading battery scheduling and carbon aware operational planning Solar and wind resources are inherently intermittent their output fluctuates with cloud cover wind speed atmospheric turbulence seasonal patterns and local terrain The proliferation of IoT and edge devices spanning smart meters inverters anemometers pyranometers weather stations and grid interface sensors has created an unprecedented volume of real time operational data that conventional forecasting pipelines are ill equipped to exploit fully This review investigates how large language model LLM agents can enhance renewable energy forecasting by integrating heterogeneous sensor streams weather API data historical generation records grid constraints and contextual reasoning into unified decision support workflows We survey classical forecasting methods statistical time series models deep learning architectures physics hybrid approaches and emerging LLM agent frameworks for explanation uncertainty communication and operator guidance A six layer taxonomy is proposed covering data acquisition preprocessing feature engineering model inference uncertainty estimation and natural language reporting The review identifies twelve open challenges spanning real time deployment model drift under distribution shift uncertainty quantification hallucination control in LLM agents interoperability of edge hardware and integration with energy management systems The paper concludes by recommending a research agenda centred on open benchmarks physics informed LLM grounding and federated forecasting architectures

Reliably knowing when a language model is correct is almost as important as being correct. We introduce prover-verifier deliberation (PVD), an inference-time protocol grounded in interactive proof theory, as a mechanism for selective prediction: the protocol produces both an answer and a structured confidence verdict, allowing a system to report high-confidence answers while abstaining on uncertain cases. In each dialogue, a prover defends a candidate answer through checkable sub-claims while a verifier issues targeted challenges and returns Accept, Challenge, or Reject. Because frozen language models are imperfect provers and verifiers operating over a noisy channel, formal soundness and completeness guarantees do not transfer; instead, we characterize the protocol empirically through its coverage-precision behavior. Our main experiment uses Claude Sonnet 4.6 as prover and Claude Haiku 4.5 as verifier on GPQA Diamond. Questions accepted with no answer revision, which we call Accept + No Change (ANC), are reported as the high-confidence subset; we evaluate this subset by its precision and coverage. ANC separates reliable from unreliable answers, yielding a $\sim$30pp HC-Prec gap over the non-ANC complement. Robustness experiments with GPT and Gemini pairings show that high HC-Prec can transfer across model families, while verifier strictness and domain competence largely determine the size of the selection gap. On Humanity's Last Exam, weaker prover-verifier pairings can collapse or invert the ANC signal, illustrating a practical failure mode when the verifier operates outside its effective region. Comparisons with self-consistency, universal self-consistency, multi-agent debate, and Reflexion suggest that prover-verifier deliberation supplies a distinct argument-defensibility signal for selective prediction.

We study inference-time alignment for diffusion-based generative models, aiming to steer a base model toward high-reward outputs without updating its weights. Recent Sequential Monte Carlo (SMC)-based steering methods approximate reward-tilted target distributions in a principled way, but their proposals remain largely tied to the base sampler. Since reward information is mainly used after propagation through particle reweighting and resampling, these methods can require large particle budgets and suffer from weight degeneracy and high-variance estimates. One way to reduce variance and improve particle efficiency is to iteratively learn twisting functions that provide look-ahead guidance, as in twisted SMC. However, existing learnable twisting methods are developed mainly for classical sequential inference and can be unstable when applied to diffusion-based alignment with high-dimensional state spaces and terminal, noisy, or black-box rewards. We propose Trust-Region Iterative Twisted Sequential Monte Carlo (TRI-TSMC), a trust-region framework for learning twisting functions in SMC-based inference-time alignment. Each iteration computes an exact KL-constrained update in path space, which admits a closed-form solution by tempered importance reweighting, and projects this target back to the parameterized twisted family by weighted maximum likelihood. Theoretically, we formalize the value-function interpretation of the optimal twisting function and show that it yields a zero-variance sampler. We prove that the trust-region update follows an escort path toward the target distribution, that the weighted maximum-likelihood update is a forward-KL projection, and that the path reduces residual importance-weight variance. Empirically, TRI-TSMC improves primary alignment objectives on discrete diffusion text generation and text-to-image generation under matched inference-time budgets.

Radiology reports remain the primary mechanism by which imaging findings are communicated to clinical teams. However, much of the structured information behind these reports, including measurements, image evidence, prior comparisons, lesion identity, uncertainty, and terminology, often remains trapped in free text or fragmented across picture archiving and communication systems, radiology information systems, reporting workstations, worksheets, advanced visualization tools, and electronic health records. This paper proposes a human-supervised, evidence-linked reference architecture for structured radiology reporting. The framework combines exam-specific templates, speech-to-structure processing, measurement and segmentation capture, controlled AI-assisted drafting, and standards-based interoperability using DICOM, DICOM Structured Reporting, DICOM Segmentation, HL7 FHIR, RadLex, SNOMED CT, LOINC, and UCUM. The system is positioned not as an autonomous report generator, but as a structured intelligence layer for enterprise imaging that supports reviewed reporting, longitudinal comparison, clinical data reuse, governance, and integration with PACS, RIS, EHR, analytics, and registry workflows. The paper also discusses modality-specific deployment considerations, clinical safety risks, validation requirements, cybersecurity, privacy, quality management, and regulatory boundaries for AI-assisted radiology reporting systems.

Long-term conversational memory is a retrieval workload classical IR was not built for: the index grows during the query stream, query types shift intra-session, and the latency budget per retrieval is sub-10 ms. Lucene-class engines treat the index as static and the query as stateless, leaving the workload's structure unexploited. AgentIR treats fusion as a per-query decision along two axes: which fusion to apply (BM25, Dense, RRF, or agent-aware RRF), and whether the ~52 ms dense channel is worth running at all. The second axis is a confidence-triggered cascade router that decides from the BM25 top-k margin alone and re-tunes across workloads without retraining. On LongMemEval (n=500), where the dense channel does add information, the cascade skips 63% of queries at parity LLM-judged accuracy (2.67x faster under two judges, paired bootstrap p>=0.88); per-qtype thresholds extend this to 5.76x under 5-fold cross-validation. On LoCoMo (n=1,982), where BM25 alone is already the strongest single system, the same trigger auto-tunes to a 100% skip rate (132x faster, +0.089 Hit@5). Capacity on a shared 8-core VM rises from ~154 to ~1,400 concurrent agents (9x). Underneath the cascade, a time-partitioned index does O(log 1/epsilon) work independent of corpus size: 1234x corpus growth costs only 3.6x latency, ending in 1769x over sequential at sub-100 us p50 on 5M records. At parity quality with Lucene on 9 BEIR datasets up to 8.8M docs, the substrate runs 10x geo-mean over Pyserini 8T and 11x over PISA-1T BlockMax-WAND; an A100 reaches 1.8-39x over Pyserini 8T; chunked index build sustains 56.8K docs/sec on MS MARCO. Three subtle BM25/GPU correctness pitfalls that silently regress nDCG@10 by 6-8x are documented and fixed; post-fix CPU and GPU agree within 0.0002 nDCG@10 on all eight datasets that fit a single A100.