Forecasting emotion in conversation is increasingly framed as predicting the next emotional state's value. We show this targets a hard quantity in a state-observed testbed and propose a learnable, deployable reformulation instead. On valence-arousal-dominance trajectories from a dyadic corpus, repeating a speaker's own previous state sets a persistence baseline that no frozen-encoder correction meaningfully improves upon and a state-space filter exceeds; the floor requires the historical state to be observed, since an estimated anchor collapses persistence below a direct forecaster. What survives is the timing of a departure from persistence, learnable on three of four corpora and near chance on the highest-base-rate corpus, a negative result we report. We recast the task as selective departure forecasting: a lead-time-weighted metric with a leakage protocol, and commitment policies whose best choice switches with the departure rate. A dialogue-level conformal risk control bounds the expected loss of committed predictions, meeting the target in all twelve configurations tested, though the guarantee is bought through abstention, with coverage and alarm recall falling sharply on the hardest corpus. One tested language model's confidence fails as a ranking signal where another succeeds, a model-dependent failure not yet attributable to family versus scale. Code and result logs are available from the corresponding author upon request.
Comparisons in conversational emotion-shift forecasting are easily distorted when a learned model receives rich features but its transition baseline is given a noisy or unavailable current-emotion label, when hyperparameter budgets differ, or when target-turn information enters the representation. We present an information-matched, leakage-safe auditing protocol that closes these gaps with an explicit commitment point, executable causality tests, equal four-configuration validation searches, and seed-dialogue-aware inference. Applying it to six causal model families on four text corpora and two feature-source sensitivity configurations shows that a deployable train-only current-emotion classifier feeding the transition baseline yields a different verdict than an oracle diagnostic where both sides receive gold current emotion. Under the deployable condition, learned models improve over the predicted-label baseline on all four primary corpora; DailyDialog gains are large and survive a 72-comparison Holm correction (best ROC-AUC gain = +0.167, adjusted p=.0144). Under matched gold information, the baseline remains competitive on IEMOCAP and stronger on MELD, while every model retains a large DailyDialog advantage. We additionally report precision-recall AUC, calibration, a strictly utterance-local feature control, same-person self-shift forecasting, and a future-access dose-response test. An eight-item leakage checklist accompanies the protocol. Code, raw seed predictions, and reporting artifacts are available from the corresponding author upon request.
Recycling a target model's own predictions offers a training-free source of proposals for speculative decoding. In vision-language models, however, existing recycling indexes predictions by the last token alone, conflating predictions computed under different images and prefixes; we find that such reuse fails most often at image-sensitive decoding states. Longer keys reduce this aliasing but match fewer states. We introduce , which stores target predictions under one-, two-, and three-token keys: contextual memories live only for the current request and are thus implicitly conditioned on its image, while token-level memory persists across requests to retain coverage. At each lookup, entropy-confidence weighting fuses the retrieved predictions according to their key specificity and concentration. Best-first search composes the fused transitions into a candidate tree, supplemented by suffix-matched continuations, and one target pass verifies the tree under greedy decoding. Without training or an auxiliary draft model, \ achieves mean speedups over autoregressive decoding of 1.66and 1.56across seven single-turn benchmarks on Qwen2.5-VL-7B and LLaVA-1.6-7B, exceeding the strongest training-free baseline, Token Recycling (1.51and 1.37). Our implementation is available in https://github.com/Senmo996/RECAP-official-code.
A learned cost model that predicts query runtimes is trained on measurements from one machine. When the database moves to other hardware, its plans stay the same but their runtimes change, each by its own factor. We ask how to adapt such a model from k queries measured on the new machine: which queries to run, how to use their runtimes, and how many are enough. On PostgreSQL we timed the same plans of 1,162 TPC-H and TPC-DS queries in six environments: a laptop and a server, each with all its processors or pinned to one, the server also pinned to two, and the laptop also with a small buffer pool. No single factor describes the slowdown between two environments: most of its variance lies between query templates, so a speed factor stops improving after a few queries. A correction that fits separate lines for serial and parallel plans and adds an offset for every measured template brings the moved model, with one measured query per template, to an error (geometric mean over templates of the median Q-error) of 1.10 on TPC-H and 1.19 on TPC-DS, close to the same correction on a model trained on the new machine; fine-tuning a DACE-style neural cost model on the same queries gains at most 2%. On templates never measured, no correction we tried is on average more than 2% better than the speed factor. We then fixed hypotheses and timed new data. On TPC-H at ten times the scale the errors follow these results, but two of five repeated tests miss. On the Join Order Benchmark, whose queries differ widely within a template, the model errs by 2.36 even when retrained on the new machine, no correction gains much, and only the test on new templates holds; boosted trees and a neural model fit these queries better, and correcting the neural model works as well as fine-tuning it. We report every test as registered.
Biological discovery requires using molecular knowledge and new measurements to infer responses beyond the conditions already observed. We introduce CellCanon, the first agentic cellular world model to unify single-cell expression prediction, evidence-grounded explanation, and observation-driven updating. Its predictive core reuses condition-level biological evidence across gene-specific numerical readouts to guide single-cell expression generation. A shared world state links predictions, acquired measurements, and source-linked claims, supporting explanation and revision. On 339 held-out SciPlex3 conditions across three cell lines, CellCanon improves differential-expression AUPRC and Macro-F1 over the best baseline for each metric by 63.3% and 17.3%, respectively, and direction Macro-F1 by 7.2%. Relative to gene-wise reasoning on matched drug-cell cohorts, measured language-model token use falls by 99.77% for Predict and 98.41% for Explain. On the source-grounded benchmark, Explain correctly answers all answerable questions with supporting proofs and identifies all insufficient-evidence cases. In simulated experimental feedback, Update uses one measured source context per drug to exceed the best fine-tuned baselines in differential-expression AUPRC and Macro-F1 without changing model weights. Together, these capabilities support a cumulative account of cellular responses, with predictions and biological judgments that can be revised as evidence accumulates.
Confidence estimates guide human review in LLM-built knowledge graphs, but better error prediction can yield little improvement in an audit. We study this gap through repairability, selection headroom and label validity. Modelling non-adaptive auditing as budgeted covering, we characterise the minimal decision-relevant error statistics and bound the regret of scalar-preserving first-order planning by a structural mass Δ^{cl}. We simulate ideal repair against benchmark-aligned labels for five extraction configurations and two cross-document graphs on Re-DocRED, followed by frozen confirmation on test and replication on 500 hash-selected training documents. On the held-out cross-document graphs, an error-rate fan-out queue leaves 3.98-4.74 percentage points more errors than a joint model with the same actions. By comparison, opening shared actions for that joint planner reduces residual errors by 0.89-1.43 points relative to its own fact-only plans, in point estimates. Equivalence between joint and independent-block models remains inconclusive. Generic LLM verification improves three of four model-budget configurations on each held-out set. LLM adjudication identifies possible benchmark omissions.
Surrogate modeling of transport requires operators that capture the combined influence of spatially distributed sources and Dirichlet and Neumann boundary conditions, together with training datasets that can be generated at low computational cost. This work presents MIGONet, a multiple-input graph neural operator, and a discrete method of manufactured solutions (MMS) that trains it without solving a partial differential equation, using steady heat conduction as the model problem. In the discrete MMS procedure, temperature fields are prescribed a priori, and the source terms and Dirichlet or Neumann boundary data are evaluated with the same discrete operators as the reference finite-difference method (FDM), so every training sample satisfies the discrete governing equations by construction. Because the boundary data are derived from each field, a Dirichlet or Neumann condition can be assigned to every boundary segment of every sample, and the boundary specification becomes an operator input. MIGONet supplies representations of the complete source field and boundary data to every node before local message passing. On the rectangular domain, it is compared with a version without the complete-source encoder (MIGONet-local) and with the multiple-input operator network MIONet, using the same training data. MIGONet achieves mean relative temperature errors of 10.31% and 10.64% for curved-wire and Gaussian-spot heating, compared with 45.97% and 48.34% for MIONet and 72.25% and 67.82% for MIGONet-local. For 10,000 samples, discrete MMS reduces data-generation time from 486.65 s to 3.15 s relative to matched iterative FDM labeling. Separately trained MIGONet models are also evaluated on L-shaped, disc-shaped, and ring-shaped domains.
When does an AI assistant improve human decision-making, and when does it hurt? We argue that the answer depends not just on AI accuracy but also on how much its evidence overlaps with the human's and how the human combines the two. A worker makes a binary decision by observing her own signal and receiving a recommendation from an AI. We develop a human behavioral model with two primitives: her awareness of the overlap and her trust in the AI. Our analysis gives a closed-form phase diagram over AI capability and overlap, with regions of complementarity, impairment, and automation. A worker who neglects the overlap (correlation neglect) can be made less accurate by a more capable AI, double-counting the shared evidence. Trust reshapes the regions - undertrust may compensate for correlation neglect, so that the trust level maximizing accuracy depends on the deployment parameters. Empirically, we identify the two model parameters by pairing the same workers with different AIs, and break the accuracy loss into a correlation-neglect cost and an undertrust cost, one tied to reducing overlap on the AI side, the other to calibrating trust on the human side. On ChatBench, AI assistance raises accuracy by 21 points, yet workers apply only about a quarter of the Bayes-rational weight to AI information; GPT-4o falls in the automation region and Llama-3.1-8b stays at the complementarity boundary; and calibrating trust recovers about 2.6 points for GPT-4o but almost nothing for Llama-3.1-8b.
Temporal-difference (TD) learning underlies many prediction and control algorithms, but can diverge when combined with off-policy learning, bootstrapping, and function approximation. TD with Regularized Corrections (TDRC) is theoretically sound and practically simple, but existing convergence theory covers only the one-step case under i.i.d. sampling, while the Huber Bellman-error formulation has not been extended to TDRC or eligibility traces. In this work, we establish convergence of constrained TDRC(λ) with diminishing stepsizes and non-i.i.d. off-policy data, for state-dependent and history-dependent λ-parameters. We also propose TDRC(λ)-Huber, which preserves the TDRC(λ) fixed point while reducing sensitivity to large Bellman errors. For both algorithms, we apply an ODE-based weak-convergence method from stochastic approximation theory. For the nonlinear mean dynamics of TDRC(λ)-Huber, we give an integral small-gain condition that guarantees uniqueness and global asymptotic stability of the target equilibrium. We then establish convergence of the constrained algorithm, derive computable sufficient conditions for convergence, and identify a regularization-based convergence mechanism shared by linear TDRC(λ) and TDRC(λ)-Huber. When the mean TD matrix is positive stable, both constrained algorithms converge under sufficiently strong regularization without requiring positive definiteness.
Machine unlearning aims to remove a user's data from a trained model after a deletion request. Existing methods are majorly evaluated on three axes - how well they forget, how much capability they retain, and how robust they are to membership inference attacks, but not on whether the provider can offer the user any evidence that the deletion occurred. To fill this gap, we introduce auditability as a fourth axis, defined as the separation between requested (forget set) and retained (retain set) data under a statistic the provider computes. We show that across eleven unlearning methods on TOFU, MUSE and RWKU no method satisfies all four axes, with different methods falling short on different axes. To address this, we propose a two stage method that can be used on top of any unlearning method, in the the first stage we propose {GapWiden} which uses a hinge loss to increase the difference between audit statistics of forget set and retain set. In the second stage we propose {GapVeil}, an inference time rejection-sampler which uses the base model as the drafter and re-samples based on the audit statistics of the drafted response. We observe that using our method, auditability and leakage reach oracle parity for every unlearning method, and forget performance converges to the oracle; retain performance improves for nine of eleven methods but remains short of parity. Our results show that evidence of deletion or auditability can be constructed by design, and that treating it as an explicit requirement changes the viability of existing unlearning methods.
Large language model (LLM) shopping assistants now recommend products in multi-turn conversations. However, reliable evaluation faces Data scarcity and Difficulty Standardizing Criteria. To address these obstacles, we introduce CARTBench for CART (Complex and Ambiguous Requests with Timelines). For Data scarcity, we curate cases from real commercial shopping needs across 17 retail domains, pairing timestamped dialogue with 1,603 product records. For Difficulty Standardizing Criteria, we develop case-specific rubrics guided by Ideal Response Principles, distilled from excellent replies manually selected from successfully completed orders. These principles also guide Principle-Guided CART Prompting, our structured recommendation strategy. The experimental benchmark contains 1,328 expert-annotated criteria across Core, Excellence, and Safety. To answer the CART question, we evaluate 7 flagship commercial API models, including GPT-6.1-Sol and Claude-Opus-5, and 35B+ parameters open-weight models. In the first full-suite evaluation under the frozen rubrics, 3 of 10 models exceed 90% on Core, while none exceeds 45% on Excellence. Therefore, we identify an understanding-to-execution gap between required behavior and high-quality purchase advice. To examine these limitations, we diagnose shared failures in product evidence and unresolved needs, linked to changing requirements and decisive unknowns. Code and data are available at https://github.com/Da1yuqin/CARTBench.
Various land-use models characterize land-use patterns and facilitate scenario planning. However, existing models are largely predictive, forecasting transitions based on numeric data without synthesizing new land-use maps from simple text prompts. Meanwhile, generative AI in urban systems remains unimodal, weakly controllable, and evaluated mostly by generic image quality rather than domain-specific metrics. To address these gaps, we introduce DualControlNet, a generative land-use model that jointly synthesizes high-resolution land-use maps and satellite imagery, addressing multimodality, controllability, and verification challenges. Trained on over 40,000 paired tiles from Orlando and Chicago, DualControlNet is evaluated using land-use, spatial consistency, and image quality metrics. Relative to unimodal baselines, DualControlNet improves land-use diversity consistency by 62.8% and building coverage ratio consistency by 16.3%, more than doubles cross-modal spatial consistency, and triples visual quality. Scenario-based experiments demonstrate that DualControlNet supports sustainable development through multidimensional controls, increasing density and diversity, and responding to text prompts like "high-density development." Overall, this work advances land-use modeling via a multimodal, controllable, and verifiable generative framework, lowering technical barriers and enabling intuitive exploration of urban futures.
Prenatal ultrasound biometry must return a head circumference (HC) and refuse when the image cannot support one. Selective prediction trades coverage against error, but its finite-sample guarantees bound the rate of gross errors, not their magnitude. We audit a tool-grounded biometry pipeline on two separately pre-registered four-way splits of the 999 labelled HC18 images, retraining the segmenter three times on each. Calibrating the contour-completeness threshold by exact Clopper-Pearson risk control returns a certificate indistinguishable from a hand-set floor, yet the pipeline still emits a 271.3 mm error: a small, geometrically complete ring fitted to an image whose true HC is 324.6 mm. The accepted candidate scores 0.94 completeness, so the score cannot see the failure, and one outlier among 150 answers leaves the certified risk bound (α= 0.10) intact. A second, untouched split reproduces the mechanism on a different image (233.3 mm). Constraining the admissibility region---the anatomical range in which the measurand is defined---removes the family instead: raising the lower bound from 50 to 55 mm costs 0.4 coverage points and cuts the worst-case error on both splits to 11.4 and 11.1 mm, whereas suppressing it through the completeness threshold costs 19.0 and 12.1 points. At an anatomically defined 70 mm the frozen tests give 1.91 and 1.81 mm MAE.
Foundation-model predictors of perturbation response are evaluated primarily by aggregate accuracy. Using their predictions to prioritize experiments requires assessing each prediction's reliability. We audit the state-transition (ST) decoder used in STATE across three encoders (SE-600M, scGPT, and Tahoe-x1) and four cell lines under CRISPRi. The decoder's first attention layer has a low-rank output. Projecting this output onto 4-32 fixed directions estimated from training activations preserves mean per-perturbation Pearson delta correlation (PDC). This low-rank attention bottleneck bounds the decoder's reliability by distance to training responses: PDC declines as the cosine distance from a perturbation's observed response to its nearest training response increases. We propose predicted-NN-distance, the cosine distance from a predicted response to its nearest training response, as a per-prediction confidence score. The score has a mean absolute Spearman correlation of 0.81 with PDC, close to the 0.88 obtained with the observed response. Thresholds calibrated on this score transfer across encoders and cell lines, and the score remains predictive of PDC for drug perturbations.
User preferences shift as needs, context, and tastes evolve. This matters for personalized recommendation, LLM response generation, and RLHF-trained reward models, which can become miscalibrated as preferences drift. Existing non-stationary bandit methods forget uniformly across all dimensions, but real preference shifts are typically partial: some dimensions go stale while others remain valid. Neural contextual bandits accumulate confidence monotonically via the Gram matrix, with no mechanism to direct uncertainty where predictions have gone wrong. We propose PEARL (Prediction Error-Adaptive Regularized Learning), which reverses the Gram update along feature directions where predictions are surprisingly wrong, while leaving confidence intact elsewhere. We prove that (i) any non-adaptive policy suffers linear regret after a preference shift, (ii) no fixed exploration rate achieves low regret in both stationary and non-stationary regimes, and (iii) PEARL's cumulative regret scales with the number of shifted users N_s, not the total population N. On PRISM, a personalized LLM response selection benchmark (1,500 users, 21 LLMs), PEARL beats greedy by +4.0 points and the best neural bandit baseline by +2.5 points (p<0.001), while performing on par with greedy when preferences are stable. It beats sliding-window and discounted approaches by 3 to 10 points, showing that selective forgetting outperforms uniform forgetting. The selection policy is a ~100K-parameter MLP that runs in milliseconds on CPU, leaves LLM weights frozen, and works with any API-based model.
This vision paper argues that distributed learning, including federated and decentralized, should become a first-class paradigm in LLM post-training, not merely a privacy add-on. Current methods like RLHF and DPO rest on an invisible assumption: preference data should be centralized. We contend this assumption is scientifically unjustified, practically unsustainable, and ethically problematic. As LLMs deploy into healthcare, law, finance, and multilingual communities, preference data originates from populations that cannot or will not centralize. We develop a three-pillar argument grounded in (1) privacy and data sovereignty, (2) pluralism and representational equity, and (3) robustness and scalability. We formalize key technical challenges unique to this intersection, reference drift, preference heterogeneity, and the DP-utility trade-off for contrastive losses, and propose concrete algorithmic directions, and sketch a research agenda. Our analysis shows distributed post-training is not only necessary but tractable, and ignoring it risks AI that is legally noncompliant, representationally narrow, and brittle at scale.
Assessing the deployment performance of learned robot policies during training is challenging, as directly measuring task success requires costly real-world trials for each candidate checkpoint. Averaged validation loss offers a rollout-free alternative, but its checkpoint rankings can be inconsistent with deployment performance. We analyze two limitations of this criterion: it penalizes action deviations without considering whether the policy can recover from them, and gives later reference predictions equal weight even when earlier deviations may lead execution away from the reference. We propose RACE, a novel rollout-free metric for ranking checkpoints using a fixed set of held-out successful trajectories. RACE calibrates action deviations against state-conditional thresholds derived from action differences at similar states in these trajectories. We aggregate these scores through a running product, so that a substantial deviation reduces the contribution of subsequent reference predictions. Averaging the resulting cumulative scores over time and across trajectories yields the checkpoint score. In simulation, RACE outperforms baselines across validation sources, while several baselines degrade sharply when the validation set changes. On real robots, its rankings align with deployment success across the evaluated policy backbones and tasks of varying difficulty. Perturbation analyses further show that RACE distinguishes deviations with markedly different effects on task success, even when baselines respond similarly.
Firms using large language models under fixed budgets must decide which model to call and whether to retry a failed task. Because a call's charge is revealed only after completion, one long response can exhaust the remaining budget and leave later tasks unserved. Our objective is to maximize the number of tasks solved before service stops. We develop a contextual bandit policy, ADP-HT, which learns each model's task-specific success probability and constructs conservative cost estimates from coarsened response-length labels on a quantile grid with 2^K intervals. The policy weighs predicted success against these costs, adjusting the cost penalty to realized spending and the budget available per remaining task. Under calibration, local regularity, and grid-coverage conditions, we bound the excess allowance in the policy's conservative cost estimates. At a fixed sample size, increasing K reduces the discretization term but increases the estimation allowance. At the tightest tested budgets, ADP-HT solves more than twice as many MATH500 and GPQA tasks as the best implemented baseline. It achieves the highest mean cumulative score at ten of fourteen RouterBench budgets. On LiveCodeBench, it outperforms a deployed commercial router in both default and cost-saving modes at the three lowest test budgets.
Intelligent production requires systems to make decisions and organize work according to production goals and operating conditions. It also requires continuous adjustment through feedback. This paper analyzes intelligence in human production organizations from a complex adaptive systems (CAS) perspective. The analysis identifies two coupled mechanisms: individual autonomy and collective coordination. Individual autonomy enables actors to form local decisions using available information. Collective coordination aligns these decisions with shared goals and joint constraints. Learning and adaptation improve both mechanisms over time. A production-system model formalizes how local decision rules and coordination jointly form system decisions. On this basis, we propose the Recursive Autonomous Agent (RAA) framework, which combines goal-driven autonomous agents with a recursive organizational structure. Goal contracts define local objectives and operating boundaries, while capability contracts support the selection and composition of production units. Common interfaces allow coordinated groups to participate as units in larger organizations, connecting equipment, production lines, and factories. RAA thus provides an engineering approach to next-generation intelligent manufacturing, translating the requirements of intelligence, openness, and system-wide integration into mechanisms for autonomous and coordinated production.
