new

Get trending papers in your email inbox!

Subscribe

Daily Papers

byAK and the research community

Nov 19

Channel Selection for Mixture-of-Experts Decoding Needs No Calibration Data

Mixture-of-Experts (MoE) models are bottlenecked by loading routed-expert weights from memory during interactive decoding at small batch sizes. Because most intermediate channels in a routed expert contribute little to any single token, channel selectors predict which channels matter and read only their weights, typically keeping channels above a threshold calibrated on offline data. We show that channel selection needs no calibration data. First, a calibrated threshold does not fix how much a selector reads, because the number of channels passing it varies across inputs; Prox with thresholds calibrated for an 18% read budget actually reads only 12.9% on average across our benchmarks. Second, top-k keeps each token's k highest-scoring channels and needs no calibration data. At the calibrated budget, a selector's top-k version matches or beats its threshold version on the same score in 40 of 43 comparisons at 13% and 18% budgets, including all 23 on generation tasks. The top-k versions also avoid the repetition-loop collapse that calibrated thresholds cause on Gemma 4. Calibrating a threshold thus provides neither of the two things selectors use it for: a known read budget and a consistent accuracy gain at that budget. We furthermore introduce a reader built entirely from the model's weights, which screens channels with a ternary sketch and recovers the skipped ones with a low-rank map. Across three post-trained MoE models, our reader retains 98.1% of full-model accuracy at 17.8% of routed-expert reads, with Triton kernels running MoE blocks 1.60 to 1.90x faster than optimized vLLM at batch one and up to 2.69x at larger batches on the same three models.

kaist-ai KAIST AI
·
Oct 26 6

SERUM: State Extraction and Refinement for User Modeling

Agentic assistants capable of proactive, personalized interactions require structured models of user intent and workflow. However, building these models from raw, unstructured screen activity remains an open challenge. We present SERUM, a multi-pass framework that extracts finite-state behavioral models directly from unstructured egocentric video using hierarchical VLM annotation. Processing screen recordings through a sliding window, SERUM alternates between activity-recognition and intent-inference passes, with each pass refining labels using accumulated prior context to reduce hallucination and temporal conflation seen in single-pass annotation. Synonymous states are then merged via sentence embeddings and human-calibrated thresholds into a compact, coherent taxonomy. We evaluate behavioral structure by fitting first-order Markov models over the resulting label sequences (both actions and intents) and measuring predictive accuracy against frequency baselines. Across 61 egocentric videos in four domains (coding, cooking, physical activities, and daily life), we find: (1) iterative label refinement converges to a stable state vocabulary, which we term schematic equilibrium, after several passes; (2) normalized Markov models achieve substantially lower perplexity and higher action predictions than frequency baselines, with the largest gains on structured tasks like coding; and (3) human annotators rate final-pass labels as accurate and meaningfully improved over first-pass labels. To our knowledge, SERUM is the first system to produce interpretable process models from unstructured egocentric screen video without manual annotation, opening a scalable pathway for user modeling and behavioral understanding in the wild. Our demo, code, and results are publicly available

  • 5 authors
·
Jul 30

When World Models Dream Wrong: Physical-Conditioned Adversarial Attacks against World Models

Generative world models (WMs) are increasingly used to synthesize controllable, sensor-conditioned driving videos, yet their reliance on physical priors exposes novel attack surfaces. In this paper, we present Physical-Conditioned World Model Attack (PhysCond-WMA), the first white-box world model attack that perturbs physical-condition channels, such as HDMap embeddings and 3D-box features, to induce semantic, logic, or decision-level distortion while preserving perceptual fidelity. PhysCond-WMA is optimized in two stages: (1) a quality-preserving guidance stage that constrains reverse-diffusion loss below a calibrated threshold, and (2) a momentum-guided denoising stage that accumulates target-aligned gradients along the denoising trajectory for stable, temporally coherent semantic shifts. Extensive experimental results demonstrate that our approach remains effective while increasing FID by about 9% on average and FVD by about 3.9% on average. Under the targeted attack setting, the attack success rate (ASR) reaches 0.55. Downstream studies further show tangible risk, which using attacked videos for training decreases 3D detection performance by about 4%, and worsens open-loop planning performance by about 20%. These findings has for the first time revealed and quantified security vulnerabilities in generative world models, driving more comprehensive security checkers.

  • 7 authors
·
Feb 21

L-RAG: Balancing Context and Retrieval with Entropy-Based Lazy Loading

Retrieval-Augmented Generation (RAG) has emerged as the predominant paradigm for grounding Large Language Model outputs in factual knowledge, effectively mitigating hallucinations. However, conventional RAG systems operate under a "retrieve-always" assumption, querying vector databases for every input regardless of query complexity. This static approach incurs substantial computational overhead and inference latency, particularly problematic for high-throughput production deployments. We introduce L-RAG (Lazy Retrieval-Augmented Generation), an adaptive framework that implements hierarchical context management through entropy-based gating. L-RAG employs a two-tier architecture: queries are first processed with a compact document summary, and expensive chunk retrieval is triggered only when the model's predictive entropy exceeds a calibrated threshold, signaling genuine uncertainty. Through experiments on SQuAD 2.0 (N=500) using the Phi-2 model, we demonstrate that L-RAG provides a tunable accuracy-efficiency trade-off: at a conservative threshold (tau=0.5), L-RAG achieves 78.2% accuracy, matching Standard RAG (77.8%), with 8% retrieval reduction; at a balanced threshold (tau=1.0), retrieval reduction increases to 26% with modest accuracy trade-off (76.0%). Latency analysis shows that L-RAG saves 80-210ms per query when retrieval latency exceeds 500ms. Analysis of entropy distributions reveals statistically significant separation (p < 0.001) between correct predictions (H=1.72) and errors (H=2.20), validating entropy as a reliable uncertainty signal. L-RAG offers a practical, training-free approach toward more efficient RAG deployment, providing system architects with a configurable knob to balance accuracy and throughput requirements.

  • 1 authors
·
Jan 9

Human Values in a Single Sentence: Moral Presence, Hierarchies, and Transformer Ensembles on the Schwartz Continuum

We study sentence-level identification of the 19 values in the Schwartz motivational continuum as a concrete formulation of human value detection in text. The setting - out-of-context sentences from news and political manifestos - features sparse moral cues and severe class imbalance. This combination makes fine-grained sentence-level value detection intrinsically difficult, even for strong modern neural models. We first operationalize a binary moral presence task ("does any value appear?") and show that it is learnable from single sentences (positive-class F1 approx 0.74 with calibrated thresholds). We then compare a presence-gated hierarchy to a direct multi-label classifier under matched compute, both based on DeBERTa-base and augmented with lightweight signals (prior-sentence context, LIWC-22/eMFD/MJD lexica, and topic features). The hierarchy does not outperform direct prediction, indicating that gate recall limits downstream gains. We also benchmark instruction-tuned LLMs - Gemma 2 9B, Llama 3.1 8B, Mistral 8B, and Qwen 2.5 7B - in zero-/few-shot and QLoRA setups and build simple ensembles; a soft-vote supervised ensemble reaches macro-F1 0.332, significantly surpassing the best single supervised model and exceeding prior English-only baselines. Overall, in this scenario, lightweight signals and small ensembles yield the most reliable improvements, while hierarchical gating offers limited benefit. We argue that, under an 8 GB single-GPU constraint and at the 7-9B scale, carefully tuned supervised encoders remain a strong and compute-efficient baseline for structured human value detection, and we outline how richer value structure and sentence-in-document context could further improve performance.

  • 2 authors
·
Jan 20

Label-Free Detection of Governance Evidence Degradation in Risk Decision Systems

Risk decision systems in fraud detection and credit scoring operate under structural label absence: ground truth arrives weeks to months after decisions are made. During this blind period, model performance may degrade silently, eroding the governance evidence that justifies automated decisions. Existing drift detection methods either require labels (supervised detectors) or detect statistical change without distinguishing harmful degradation from benign distributional evolution (unsupervised detectors). No existing framework integrates drift detection with governance evidence assessment and operational response. This paper presents a label-free governance monitoring extension to the Governance Drift Toolkit that produces governance alerts rather than statistical alarms. The monitoring architecture applies composite multi-proxy monitoring across four proxy monitors (score distribution, feature drift, prediction entropy, confidence distribution), with governance-calibrated thresholds. Empirical evaluation on the Lending Club credit scoring dataset (1.37M loans, 11 years) demonstrates three findings. First, raw proxy metrics (Feature PSI delta up to 1.84, Score PSI delta up to 0.92) distinguish injected covariate degradation from natural temporal drift in an offline evaluation setting. Second, pure concept drift in P(Y|X) produces exactly zero delta across all proxy metrics in all windows, confirming the irreducible blind spot of label-free monitoring as a structural verification. Third, the composite score provides monotonic severity progression as more monitors trigger (0.583 to 0.833 to 1.000), enabling graduated governance response. Cross-domain comparison with IEEE-CIS fraud detection results shows the detectable/undetectable boundary is consistent across both domains. The toolkit and evaluation code are available as open-source artifacts.

  • 1 authors
·
Apr 19

Decomposing the Generalization Gap in PROTAC Activity Prediction: Variance Attribution and the Inter-Laboratory Ceiling

Machine-learning predictors of biochemical activity often exhibit large random-split-to-leave-one-target-out generalisation gaps that have been documented but not decomposed. We frame this as an evaluation-science question and use targeted protein degradation as the empirical test bed. PROTACs (proteolysis-targeting chimeras) are heterobifunctional small molecules that induce targeted protein degradation, with more than forty candidates currently in clinical trials; published predictors report AUROC of 0.85 to 0.91 under random-split cross-validation, while the leave-one-target-out (LOTO) protocol of Ribes et al. reduces performance to approximately 0.67. Random splits reward within-target interpolation, whereas LOTO measures the novel-target prediction that de-novo design depends on. We decompose this gap and identify inter-laboratory measurement variance as the dominant component, anchored by a within-target cross-laboratory cascade bounding the inter-laboratory contribution at 0.124 AUROC, well above the 0.05 contribution from binarisation-threshold choice. Across eight published architectures and ESM-2 protein language models up to 3B parameters, LOTO AUROC plateaus near 0.67, with a comparable plateau under SMILES-level deduplication; a 21-dimensional 2000-trial hyperparameter optimisation cannot break this ceiling, and the rank-1 single-seed configuration regresses by 0.161 AUROC under multi-seed validation, matching a closed-form selection-bias prediction (Bailey and Lopez de Prado, 2014). Few-shot k=5 stratified per-target retraining combined with ADMET features lifts 65-target LOTO AUROC from 0.668 to 0.7050, and post-hoc Platt scaling recovers raw output to within the 0.05 well-calibrated threshold. We release PROTAC-Bench (10,748 measurements, 173 targets, 65 LOTO folds), the variance-decomposition framework, the per-target calibration protocol, and the evaluation code.

  • 3 authors
·
May 11

When Should a Network Emit Geometry, and When Should It Detect It? Readout, Reconciliation, and Representation in Floorplan Vectorization

A network trained to recover the walls, openings, and rooms of a rasterized floorplan can produce its output in two ways: by emitting the geometry as an autoregressive coordinate sequence, or by detecting it on dense junction and centerline heatmaps and assembling a graph. We compare the two readouts on the same trained network. On real scans (CubiCasa5K) detection is better on every wall measure (+2.7 wall F1 at tolerance 0.05, +5.1 at 0.015; paired bootstrap intervals exclude zero), and reading an opening heatmap the decoder never used raises opening F1 by 2.6x without retraining. Within real scans the readout's advantage grows with plan size and reverses on small plans; on clean vector renders sequence decoding is better by 5 to 8 points where its training covered the render style, while under full domain shift the readout, given calibrated thresholds, stays ahead; neither ink density nor plan size explains the reversal. With matched data and recipe, a room-centric system with a reconciliation step and a wall-first sequence model reach comparable wall quality, so the output representation matters less than is usually assumed. A prior from the other family helps at the output but not at the input: deterministic fusion of the two outputs raises wall F1 by 7 points, whereas conditioning one model on the other's output gives no gain in three forms, including two ground-truth-content controls. We also provide an edit-cost metric that scores a draft by the human work needed to correct it, corrected CubiCasa5K annotations, and ResPlan-FP, a CC BY 4.0 benchmark of 16,998 plans with frozen splits and three baseline tracks. Code, the benchmark, and the corrected annotations are available at https://github.com/Cyprinus12138/fpvec-lab

  • 1 authors
·
Aug 25

Agent Retrieval Bench: Evaluating Repository Context Retrieval for Coding Agents

Modern coding agents are usually evaluated by whether they eventually produce a correct patch, but patch generation depends on an earlier context-acquisition stage: finding the repository files needed for the task. We introduce Agent Retrieval Bench, a file-level benchmark for this upstream retrieval problem. Samples are built from real coding-workflow signals and evaluated against frozen base-commit repositories, with relevance defined by what an agent needs next rather than direct query-file semantic similarity. The benchmark covers four positive-retrieval tasks: code2test, comment2context, trace2code, and edit2ripple; a fifth subset evaluates selective retrieval using natural evidence-backed no-gold cases and counterfactual wrong-repository controls. Agent Retrieval Bench contains 427 samples across 25 repositories: 345 positive examples, 50 natural no-gold examples, and 32 counterfactual controls. The corpus includes 308 base-commit snapshots, 392,000 files, and 7.9 million chunks. We evaluate lexical retrieval, RepoMap, open-source embeddings, selective abstention, and logged agent context selection. No single retrieval family dominates: Qwen3-Embedding-4B has the best sample-weighted MRR on positive samples, Qwen3-Embedding-8B the best Recall@20, and RepoMap the best budgeted context yield at 8K tokens, with task-level winners differing substantially. Selective thresholds calibrated with counterfactual controls do not improve selective success on natural no-gold cases, revealing a calibration gap. Logged trajectories also miss every gold file on 27-35 percent of samples. A controlled seed-intervention pilot finds that retrieval-derived initial context yields higher file F1 with less post-seed exploration than random non-gold context, while oracle gold context shows substantial remaining headroom.

  • 2 authors
·
Jul 26 2

Stochastic Deep Supervision for Gallery-Consistent Early-Exit Face Recognition

Vision Transformers (ViTs) perform strongly in face recognition (FR), yet their fixed-depth inference limits deployment in resource-constrained settings. We propose SD-ViT, a training framework that gives a usable exit at every block depth by making every depth produce discriminative embeddings, for budget-selected fixed-depth exits. Standard ViTs supervise intermediate layers only through backpropagation, so their intermediate representations are not directly optimized to form usable identity embeddings. SD-ViT adds a stochastic auxiliary signal: at each iteration a randomly selected intermediate block is passed through the same embedding head as the final layer and trained with a weighted margin-penalty softmax loss. The objective is thus stochastically sampled deep supervision through a shared classifier. Because both loss terms share that head and classifier, intermediate tokens are trained to map into the same embedding space as final-layer tokens. This matters for FR, where matching is by similarity: a probe processed at an early exit can be compared directly against full-depth gallery embeddings, with no gallery recomputation or adaptation layers. On established benchmarks, SD-ViT stays within 0.4 IJB-C points of the same backbone trained conventionally at full depth, while its early-exit probes recover over 96% of their own full-depth performance at 50% of the computational cost without modifying the reference gallery. A single verification threshold calibrated at full depth remains usable from quarter depth onward.

LLM Output Drift: Cross-Provider Validation & Mitigation for Financial Workflows

Financial institutions deploy Large Language Models (LLMs) for reconciliations, regulatory reporting, and client communications, but nondeterministic outputs (output drift) undermine auditability and trust. We quantify drift across five model architectures (7B-120B parameters) on regulated financial tasks, revealing a stark inverse relationship: smaller models (Granite-3-8B, Qwen2.5-7B) achieve 100% output consistency at T=0.0, while GPT-OSS-120B exhibits only 12.5% consistency (95% CI: 3.5-36.0%) regardless of configuration (p<0.0001, Fisher's exact test). This finding challenges conventional assumptions that larger models are universally superior for production deployment. Our contributions include: (i) a finance-calibrated deterministic test harness combining greedy decoding (T=0.0), fixed seeds, and SEC 10-K structure-aware retrieval ordering; (ii) task-specific invariant checking for RAG, JSON, and SQL outputs using finance-calibrated materiality thresholds (plus or minus 5%) and SEC citation validation; (iii) a three-tier model classification system enabling risk-appropriate deployment decisions; and (iv) an audit-ready attestation system with dual-provider validation. We evaluated five models (Qwen2.5-7B via Ollama, Granite-3-8B via IBM watsonx.ai, Llama-3.3-70B, Mistral-Medium-2505, and GPT-OSS-120B) across three regulated financial tasks. Across 480 runs (n=16 per condition), structured tasks (SQL) remain stable even at T=0.2, while RAG tasks show drift (25-75%), revealing task-dependent sensitivity. Cross-provider validation confirms deterministic behavior transfers between local and cloud deployments. We map our framework to Financial Stability Board (FSB), Bank for International Settlements (BIS), and Commodity Futures Trading Commission (CFTC) requirements, demonstrating practical pathways for compliance-ready AI deployments.

  • 2 authors
·
Nov 10, 2025

Is Conformal Factuality for RAG-based LLMs Robust? Novel Metrics and Systematic Insights

Large language models (LLMs) frequently hallucinate, limiting their reliability in knowledge-intensive applications. Retrieval-augmented generation (RAG) and conformal factuality have emerged as potential ways to address this limitation. While RAG aims to ground responses in retrieved evidence, it provides no statistical guarantee that the final output is correct. Conformal factuality filtering offers distribution-free statistical reliability by scoring and filtering atomic claims using a threshold calibrated on held-out data, however, the informativeness of the final output is not guaranteed. We systematically analyze the reliability and usefulness of conformal factuality for RAG-based LLMs across generation, scoring, calibration, robustness, and efficiency. We propose novel informativeness-aware metrics that better reflect task utility under conformal filtering. Across three benchmarks and multiple model families, we find that (i) conformal filtering suffers from low usefulness at high factuality levels due to vacuous outputs, (ii) conformal factuality guarantee is not robust to distribution shifts and distractors, highlighting the limitation that requires calibration data to closely match deployment conditions, and (iii) lightweight entailment-based verifiers match or outperform LLM-based model confidence scorers while requiring over 100times fewer FLOPs. Overall, our results expose factuality-informativeness trade-offs and fragility of conformal filtering framework under distribution shifts and distractors, highlighting the need for new approaches for reliability with robustness and usefulness as key metrics, and provide actionable guidance for building RAG pipelines that are both reliable and computationally efficient.

  • 5 authors
·
Mar 16

ObjexMT: Objective Extraction and Metacognitive Calibration for LLM-as-a-Judge under Multi-Turn Jailbreaks

LLM-as-a-Judge (LLMaaJ) now underpins scalable evaluation, yet we lack a decisive test of a judge's qualification: can it recover a conversation's latent objective and know when that inference is trustworthy? LLMs degrade under irrelevant or long context; multi-turn jailbreaks further hide goals across turns. We introduce ObjexMT, a benchmark for objective extraction and metacognition. Given a multi-turn transcript, a model must return a one-sentence base objective and a self-reported confidence. Accuracy is computed via LLM-judge semantic similarity to gold objectives, converted to binary correctness by a single human-aligned threshold calibrated once on N = 100 items (tau^*=0.61). Metacognition is evaluated with ECE, Brier, Wrong-at-High-Conf, and risk-coverage. Across gpt-4.1, claude-sonnet-4, and Qwen3-235B-A22B-FP8 on SafeMTData_Attack600, SafeMTData_1K, MHJ, and CoSafe, claude-sonnet-4 attains the best objective-extraction accuracy (0.515) and calibration (ECE 0.296; Brier 0.324); gpt-4.1 and Qwen3-235B-A22B-FP8 tie at 0.441 but are overconfident (mean confidence approx0.88 vs. accuracy approx0.44; Wrong-at-0.90 approx48-52%). Performance varies by dataset (approx0.167-0.865). ObjexMT thus supplies an actionable test for LLM judges: when objectives are not explicit, judges often misinfer them with high confidence. We recommend exposing objectives when feasible and gating decisions by confidence otherwise. Code and data at https://github.com/hyunjun1121/ObjexMT_dataset.

AIM-Intelligence AIM Intelligence
·
Aug 22, 2025

StopLoss: Teaching Agents to Predict Failure and Stop Early

Coding agents keep working on runs that will fail verification, and asking a frontier model to watch for this is expensive in its own right. We introduce StopLoss, a small external monitor that learns failure recognition from completed trajectories and their verifier outcomes. The name is inspired by the stop-loss order in finance: StopLoss ends a run early when the run is likely to fail, which saves resources. Final outcomes label earlier local windows, so a fine-tuned language model can score the most recent tool steps in a single forward pass, and a threshold calibrated on successful runs turns that score into an abort decision without changing the executor. Trained on GPT-family SWE-rebench trajectories and evaluated on SWE-bench Verified and a SWE-bench Pro subset with GPT and Claude executors, StopLoss is compared against six general-purpose judges reading identical windows. Outcome supervision lifts failure prediction from near chance to a useful signal at the first check, at one to two orders of magnitude lower cost per check than any tested judge. This translates into savings: relative to completing every task, StopLoss lowers cost per retained solve by 4.5\% on Verified while preserving 99\% of successful runs, and by up to 31\% on the costlier Pro subset, reaching 48\% with a Qwen3-8B backbone, and in all four settings its cost per retained solve is below that of every judge policy. Outcome-trained small monitors can make failure recognition cheap enough for early stopping to pay for itself.

  • 6 authors
·
Oct 17 2

Signed but Steerable: Measuring and Closing the Text-Vector Integrity Gap in Offline Retrieval-Augmented Generation

Offline retrieval-augmented generation (RAG) on edge devices is updated through signed knowledge packs that carry text and precomputed embedding vectors. A signature proves who published a pack, not that its vectors encode its text. We study this text-vector integrity gap. A verifier that recomputes vectors and accepts them within a cosine tolerance concedes a predictable share of queries to an attacker who controls the vectors: the hijack rate equals the distribution of angular retrieval margins at the tolerance (the margin law). Margins are small in five corpora and two embedding models, so an idealised tolerance at the level of floating-point noise leaves up to 5.5% of queries exposed (at most 1.2% in any corpus once exact duplicate documents are excluded), and a threshold calibrated from our CPU INT8-FP32 comparison of a small embedding model (at the empirical 0.1% quantile of honest cosines) leaves 98-100% exposed; on a Raspberry Pi 5 running a local language model, such an attack lowers correct answers from 0.75 to 0.39. Integrity audits of compressed vector stores fare no better: on a FAISS product-quantisation index over 57,638 documents, tampered vectors that all pass the audit put the attacker's document first for 25% of targets. Verifying quantisation codes instead, with ambiguity decided by the verifier, a calibrated guard band and canonicalisation, accepted every honest vector in our samples on three platforms and left the publisher no choice. We also show that common floating-point and dynamic-INT8 inference pipelines are not bit-reproducible across batch sizes or architectures, and report pre-registered negative results on abstention, context pruning and semantic screening. We release code and data.

Skywork R1V2: Multimodal Hybrid Reinforcement Learning for Reasoning

We present Skywork R1V2, a next-generation multimodal reasoning model and a major leap forward from its predecessor, Skywork R1V. At its core, R1V2 introduces a hybrid reinforcement learning paradigm that harmonizes reward-model guidance with rule-based strategies, thereby addressing the long-standing challenge of balancing sophisticated reasoning capabilities with broad generalization. To further enhance training efficiency, we propose the Selective Sample Buffer (SSB) mechanism, which effectively counters the ``Vanishing Advantages'' dilemma inherent in Group Relative Policy Optimization (GRPO) by prioritizing high-value samples throughout the optimization process. Notably, we observe that excessive reinforcement signals can induce visual hallucinations--a phenomenon we systematically monitor and mitigate through calibrated reward thresholds throughout the training process. Empirical results affirm the exceptional capability of R1V2, with benchmark-leading performances such as 62.6 on OlympiadBench, 79.0 on AIME2024, 63.6 on LiveCodeBench, and 74.0 on MMMU. These results underscore R1V2's superiority over existing open-source models and demonstrate significant progress in closing the performance gap with premier proprietary systems, including Gemini 2.5 and OpenAI o4-mini. The Skywork R1V2 model weights have been publicly released to promote openness and reproducibility https://huggingface.co/Skywork/Skywork-R1V2-38B.

Skywork Skywork
·
Apr 23, 2025 2

Verification Limits Code LLM Training

Large language models for code generation increasingly rely on synthetic data, where both problem solutions and verification tests are generated by models. While this enables scalable data creation, it introduces a previously unexplored bottleneck: the verification ceiling, in which the quality and diversity of training data are fundamentally constrained by the capabilities of synthetic verifiers. In this work, we systematically study how verification design and strategies influence model performance. We investigate (i) what we verify by analyzing the impact of test complexity and quantity: richer test suites improve code generation capabilities (on average +3 pass@1), while quantity alone yields diminishing returns, (ii) how we verify by exploring relaxed pass thresholds: rigid 100% pass criteria can be overly restrictive. By allowing for relaxed thresholds or incorporating LLM-based soft verification, we can recover valuable training data, leading to a 2-4 point improvement in pass@1 performance. However, this benefit is contingent upon the strength and diversity of the test cases used, and (iii) why verification remains necessary through controlled comparisons of formally correct versus incorrect solutions and human evaluation: retaining diverse correct solutions per problem yields consistent generalization gains. Our results show that Verification as currently practiced is too rigid, filtering out valuable diversity. But it cannot be discarded, only recalibrated. By combining calibrated verification with diverse, challenging problem-solution pairs, we outline a path to break the verification ceiling and unlock stronger code generation models.

  • 6 authors
·
Sep 25, 2025

SoftVTBench: A Safety-Aware Visuo-Tactile Benchmark for Physically Constrained Robotic Manipulation of Deformable Objects

Deformable object manipulation poses challenges beyond task completion: successful execution must also maintain safe physical interaction, holding the object stably without slip or drop while avoiding excessive deformation. However, existing manipulation benchmarks are predominantly success-oriented and rarely evaluate whether a policy remains physically safe throughout execution. We present SoftVTBench, a safety-aware visuo-tactile benchmark for physically constrained deformable object manipulation. Built in Isaac Sim with finite-element-simulated deformable objects, SoftVTBench provides multi-view RGB observations, RGB tactile sensing with marker motion, proprioception, and language instructions, and defines four matched task suites over object type (deformable vs. rigid) and variation axis (object vs. spatial). It separately reports Goal Success and Safety Success; the latter additionally requires no drop and peak deformation below a calibrated object-specific threshold, measured from policy-hidden privileged Finite Element Method (FEM) states. We implement pi0.5-based baselines under this protocol. Experiments show that success-only evaluation substantially overstates policy performance, as a large fraction of goal-completing rollouts still violate physical safety. Furthermore, incorporating tactile sensing improves Safety Success (e.g., from 21.4% to 35.6% on object-centric deformable tasks) and reduces object deformation during execution, while maintaining comparable Goal Success. SoftVTBench provides a reproducible benchmark for studying visuo-tactile deformable manipulation under physical interaction constraints.

  • 18 authors
·
Jul 4

Calibrated Selective Prediction Using Deep Ensembles for ROI-Based Thyroid Nodule Ultrasound Classification Under Dataset Shift: A Retrospective Evaluation

Background: Deep learning models can classify thyroid nodules on ultrasound, but reliable clinical decision support also requires calibrated probabilities, uncertainty estimation, and selective referral, particularly under dataset shift. Methods: We developed a calibrated deterministic five-member deep ensemble for ROI-based thyroid nodule classification and selective image-based triage. TN5000 was used for model development, five-fold cross-validation, member-wise vector-scaling calibration, and fold-specific threshold selection. TN3K served as an independent external dataset-shift evaluation. The framework used ConvNeXt-Tiny with squeeze-and-excitation attention, ensemble-mean malignancy probability, and mutual information (MI) as an ensemble-disagreement score. A three-tier policy assigned images to No-FNA suggestion, FNA recommendation, or radiologist review. Results: On pooled out-of-fold TN5000 predictions, the ensemble achieved AUC-ROC 0.9395, AP 0.9715, ECE 0.0088, and Brier score 0.0813. At 50% nominal MI retention, 7.2% of cases received a No-FNA suggestion, 39.9% an FNA recommendation, and 52.9% radiologist review, with 98.3% No-FNA NPV and 99.83% malignancy capture. On TN3K, AUC-ROC decreased to 0.7870, AP to 0.7254, ECE increased to 0.1899, and Brier score to 0.2281. The frozen TN5000 policy assigned 83.7% to review, 1.0% to No-FNA, and 15.3% to FNA recommendation. No malignant image entered the No-FNA pathway, but FNA-recommendation PPV fell to 76.6%. Conclusion: The framework showed strong internal discrimination and calibration, but limited external threshold transportability. Selective prediction may help identify images unsuitable for automated triage, but local recalibration, threshold validation, and prospective clinical evaluation are required before deployment.

  • 5 authors
·
Jul 13

OmniQuant: Omnidirectionally Calibrated Quantization for Large Language Models

Large language models (LLMs) have revolutionized natural language processing tasks. However, their practical deployment is hindered by their immense memory and computation requirements. Although recent post-training quantization (PTQ) methods are effective in reducing memory footprint and improving the computational efficiency of LLM, they hand-craft quantization parameters, which leads to low performance and fails to deal with extremely low-bit quantization. To tackle this issue, we introduce an Omnidirectionally calibrated Quantization (OmniQuant) technique for LLMs, which achieves good performance in diverse quantization settings while maintaining the computational efficiency of PTQ by efficiently optimizing various quantization parameters. OmniQuant comprises two innovative components including Learnable Weight Clipping (LWC) and Learnable Equivalent Transformation (LET). LWC modulates the extreme values of weights by optimizing the clipping threshold. Meanwhile, LET tackles activation outliers by shifting the challenge of quantization from activations to weights through a learnable equivalent transformation. Operating within a differentiable framework using block-wise error minimization, OmniQuant can optimize the quantization process efficiently for both weight-only and weight-activation quantization. For instance, the LLaMA-2 model family with the size of 7-70B can be processed with OmniQuant on a single A100-40G GPU within 1-16 hours using 128 samples. Extensive experiments validate OmniQuant's superior performance across diverse quantization configurations such as W4A4, W6A6, W4A16, W3A16, and W2A16. Additionally, OmniQuant demonstrates effectiveness in instruction-tuned models and delivers notable improvements in inference speed and memory reduction on real devices. Codes and models are available at https://github.com/OpenGVLab/OmniQuant.

  • 10 authors
·
Aug 24, 2023

From Retrieval to Typed Decisions: Calibrated System One Models from Biomedical Sentence Encoders

Typed decision models answer schema-constrained questions about a text in one forward pass and return probabilities meant to be thresholded. We ask whether biomedical sentence encoders trained for retrieval are good starting points for such models. We present SBERT2S1, which converts Sentence-Transformers encoders into bi-encoder, cross-head (C) and prior-fused residual (PFR) decision models, together with BIODECIDE, a biomedical typed-decision suite, and MEDLINE-S1, 243k training decisions derived from NLM indexing. Across six parent-retriever pairs, retrieval training improves zero-shot matching of content-bearing options. After fine-tuning, its effect depends on the head: across five pairs and three training-set sizes, retrieval training significantly helps PFR, which keeps the retrieval prior, in 10 of 15 comparisons, but helps C in one and hurts it in five. A matched grid of two heads and five training objectives shows that C outperforms PFR under every objective, and that the released RLCD recipe of open System One models trails cross-entropy by 2.5-3.0 points. The deficit stems mainly from its reward normalisation, which inflates the noisy score-function term 3.6-15-fold; an unbiased leave-one-out estimator recovers most of the gap. After temperature scaling, no objective is clearly better calibrated than cross-entropy. We release the code, the MEDLINE-S1 labels and a model.

  • 1 authors
·
Sep 30 2

JevAdvBench: A Benchmark and Black-Box Attacks for Reinforcement Learning for Calibrated Decisions Models

Models trained with reinforcement learning for calibrated decisions (RLCD), such as Jev, answer a typed question about an input, the state, with a probability, a choice, or a score, and software acts on the answer without a person reading it. Their robustness has not been measured: adversarial benchmarks score what a model generates or executes, whereas a typed model generates nothing and returns a well-formed answer even when manipulated. Measurement is also hard, because identical requests can return different answers, most available labels come from the model itself, and the API preprocesses each request out of view. Our key idea is to score each attacked decision against the model's own clean decision rather than against labels, and to read it against the change caused by an identical re-run. Building on this, we introduce JevAdvBench, to our knowledge the first adversarial benchmark for RLCD models, with 812 typed questions over 66 scenarios, and a black-box attack suite of 9,744 single-edit variants that each edit one part of a request, with billed input tokens confirming that the edit reached the model. On jev-1.13.0, rewording stays within 1.2 percentage points of the re-run baseline, and fields outside the schema never reach the model. In contrast, one unverified opinion appended to the state flips 12.1% of decisions, statistically tied with the strongest injected command (10.1%), and pushes 38% of confident answers below the 0.8 confidence threshold that routes them to human review. Applications built on RLCD models should therefore treat the state as untrusted, argued input. Project website: https://JevAdvBench.github.io/JevAdvBench/

  • 8 authors
·
Sep 24

Certification of Real Images through Calibrated Content Authentication

Generative models can synthesize high-quality inauthentic multimedia content that is already being misused at scale. We evaluate twenty deepfake detectors against ten generators released in the last four years and find accuracy decreasing over time, from near-perfect 99.5% to 76%. Adversarial perturbations further reduce every baseline detector to below 2% accuracy, effectively inverting the detector's assigned label. We argue that this unreliability reflects a fundamental ambiguity: generators can reproduce authentic content exactly (e.g., through memorization), so content alone cannot reveal the true provenance label.For this reason, content produced by a generator must admit a faithful reconstruction by that same generator, and finding such a reconstruction makes synthetic provenance plausible and authenticity plausibly deniable.We therefore propose and evaluate a detection paradigm that outputs a calibrated prediction of whether authenticity is plausibly deniable: a faithful reconstruction by any known generator establishes plausible deniability, while calibration bounds how often content from known generators fails to be reproduced. Our evaluation shows that (i) our detector can be calibrated so that at most 1% of generated content is wrongly certified, an operating point at which most baseline detectors reach near-zero recall, including the strongest with 93% accuracy; (ii) calibrating a stricter security threshold on attacked samples preserves this bound against adaptive adversaries within the evaluated bounded-perturbation attack space, whose perturbations break every baseline, but does not cover arbitrary adversarial transformations; and (iii) post-hoc verifiability is eroding, as 1,116 of 3,000 Reddit images resist reproduction by a 2022 generator, but only 55 to 79 resist reproduction by 2024 generators.

VX-One: Typed, Calibrated Decision Heads for Real-Time Perimeter Triage - An empirical study of unknown-attack honesty, cascade gray zones, and calibration under drift

Perimeter triage faces a latency/semantics trade-off: signature engines are fast and format-safe but blind to unseen attacks, while generative models are semantic but slow and produce unparseable output. In late 2026 a class of commercial System One decision models (TypeSafe Jev; Microsoft-Decision-1) began selling typed decision primitives - binary gates, closed-set choices, ordinal scores - as product APIs, with no academic treatment. We present VX-One: a small multi-head network (100K-10M parameters) whose outputs are these primitives - a calibrated gate, a choice over a closed candidate set that always includes an explicit unknown option, and an ordinal score, trained multi-task, temperature-calibrated per head (~0.7-1.7 ms p50 on CPU). We evaluate it on UNSW-NB15, CIC-IDS2017, EMBER2024 against calibrated XGBoost, Random-Forest, MLP, Kitsune, and measure a two-phase cascade (Suricata with 52,674 Emerging-Threats rules -> VX-One) over five days of CIC-IDS2017. Findings: (i) VX-One wins detection (AUROC/FPR@95%TPR) on UNSW-NB15 and under temporal drift on CIC-IDS2017 (0.711 vs 0.637), but XGBoost wins multi-class attribution on all three benchmarks; (ii) leave-one-family-out gate detects never-seen families on 9/9 UNSW folds yet drops below chance on CIC recon family, and abstention threshold transfers poorly across benchmarks; (iii) on payload-less reconstructions signature engines resolve ~0.1% of flows - gray zone ~99.9% of traffic, recovered by Phase 2 (combined recall >=0.978); (iv) calibration collapses under temporal drift (ECE 0.04->0.40) for every method while ranking partially survives. We provide a measured account of where typed calibrated decisions help perimeter triage, and where they fail.

  • 1 authors
·
Oct 20 2

Calibrating AlphaFold confidence scores for interface quality

AlphaFold confidence scores are widely used to rank protein complex predictions and to select subsets of models for downstream applications. Although these scores provide strong discriminatory power for distinguishing prediction quality, their numerical values were never intended to represent the probability that a prediction satisfies a given structural quality criterion. In this work, we calibrate commonly used confidence scores to estimate the probability that a predicted complex exceeds a chosen threshold of DockQ, an established measure of interface quality. On antibody-antigen and general multimer datasets, post-hoc isotonic calibration consistently improves agreement between predicted probabilities and observed DockQ success rates while preserving the underlying score ordering. The calibrated probabilities also provide more accurate estimates of the fraction of selected models expected to pass a given structural quality threshold. Although the mapping is fitted within a given prediction distribution and therefore transfers imperfectly across prediction domains, we show that reliable estimates are recovered from a small number of labeled target-domain samples. Because the calibration is entirely post hoc and requires no underlying AlphaFold model fine-tuning, it provides a straightforward and customizable way to convert existing model quality rankings into calibrated probabilities of passing a chosen structural quality threshold.

  • 3 authors
·
Oct 11

Benchmarking System One decision models against trained classifiers and language models for automated decision gates

Software that hands branching decisions to a model needs a declared option and a probability it can threshold. Typed decision models, also called System One models, return such probabilities without generating text, while supervised classifiers and generative language models are the established alternatives. One harness sends eight decision-model checkpoints from six families, including the hosted model Jev, and four open generative models from three developers the same semantic requests, and scores trained and zero-shot classifiers on the same workflow, intent, emotion and social-science items. With task labels, a fine-tuned DeBERTa-v3-large has the highest observed accuracy on every labeled benchmark but one. Without labels, no decision model is significantly more accurate than Jev on workflows or intents, but Gemma-4-31B matches it on workflows and exceeds it on CLINC-150 at higher cost and latency. Stated probabilities of generative models become unreadable when replies miss the key format, whereas key likelihoods avoid this but can saturate. A guaranteed 5 percent risk leaves Jev 0.528 of the intent decisions, and an in-scope threshold still accepts 0.310 of out-of-scope requests. On typed-decisions, swapping yes and no flips 50.5 answers per hundred for Jev and at least 16.8 for every generative model tested, against at most 6.5 for four fine-tuned decision checkpoints. Exposure to a benchmark's training data explains the largest lead of an open checkpoint, which vanishes on rater-labeled emotions. An intent-trained first stage escalating to Gemma-4-31B reaches that model's accuracy at about Jev's price. The results yield condition-dependent design rules for automated decision gates.

  • 2 authors
·
Oct 5

Adaptive-Shot Hybrid Quantum Anomaly Detection for Tactile Internet Security: Reliability-Aware Measurement Allocation Under Resource Constraints

Tactile Internet (TI) security analytics must balance reliable thresholded decisions with constrained computational and measurement resources. We study this tension for finite-shot hybrid quantum anomaly inference and introduce the Adaptive-Shot Variational Quantum Circuit (AS-VQC) policy. This validation-calibrated policy begins each record at 128 shots and cumulatively escalates through 256, 512, and 1024 shots only when the finite-shot anomaly score remains close to a validation-selected security threshold. The quantum scorer is evaluated as an off-path security analytics component rather than part of the haptic critical path. Using a 4,875-record CESNET-TimeSeries24-derived aggregate-flow benchmark, leakage-safe random, entity-group-disjoint, and temporal holdouts, and five trained quantum neural network (QNN) checkpoints per holdout, the primary AS-VQC-95 (beta = 0.95) policy averages 129.2, 276.9, and 131.2 shots per record, saving 87.4%, 73.0%, and 87.2% of the uniform 1024-shot baseline (Fixed-1024), respectively. The decision disagreement with analytic (exact-expectation) inference is 0.771%, 0.409%, and 0.635%, lower than both the uniform 128-shot baseline (Fixed-128) and a matched-budget shuffled-allocation control. Fixed-1024 remains more decision-stable, establishing a measurable reliability-resource trade-off rather than cost-free equivalence. A more conservative AS-VQC-99 (beta = 0.99) further reduces disagreement while using fewer than 512 average shots across all holdouts. These results show that finite quantum measurements can be treated as an inference resource and concentrated on boundary-sensitive TI-security decisions while exposing checkpoint-dependent escalation under unseen-entity conditions.

  • 4 authors
·
Oct 4

LLM Judges Verify Presence, Not Absence: Omission Blindness in AI Clinical Notes and What Recovers It

Ambient AI scribes draft clinical notes, and published audits find their dominant error is omission: information the encounter established that the note fails to record. The standard check is an LLM judge: a second model reads the note against the transcript and flags problems. We ask whether judges detect omissions. Public corpora cannot supply the answer key: their clinician reference notes and transcripts are materially discrepant. Our benchmark has 500 single-error note pairs from audited fact sheets, 298 with a named fact certainly absent and 202 added-or-altered controls. Across eight judge designs, paired discrimination (the flawed note below its clean twin, 0.5 a coin flip) reads 0.79-0.94 on added or altered content and 0.50-0.63 on omissions. On single notes, no design flags omissions reliably more often than perfect notes. Wording changes, voting and GEPA prompt optimisation move the operating point without creating usable detection. Restructuring the task recovers it: list the facts the transcript establishes, then check the note for each. Two methods reach it independently and trade off: a per-fact pipeline, and a GEPA-evolved prompt doing the same in one call. The pipeline's flags name the missing fact and its severity at 2.7% false alarms. The single call detects more (36.9% against 24.6%, p=0.002) at 6.2% false alarms and a tenth of the cost per note. A physician author validated 70 items and, where the two routes disagree, sided with the pipeline on 10 of 10 (p=0.002). A second clinician, not an author, graded the severity rubric blind and agrees to within a grade. On real vendor notes from a companion census no benchmark threshold transfers, but the re-calibrated single call detects more than the best of the eight at half its false-alarm rate. Omissions whose fact is restated elsewhere defeat both routes. We release the benchmark, prompts and judgements.

  • 4 authors
·
Aug 30

Beyond the Reranker: Do RAG Retrieval Enhancements Help Once a Strong Reranker Is Present?

Retrieval-augmented generation (RAG) is routinely extended with methods meant to improve retrieval: query expansion, hierarchical and cross-document summarization, graph-based expansion, per-query routing, rank fusion, and corrective re-retrieval. The benefits reported for these methods come almost exclusively from homogeneous corpora, predominantly Wikipedia prose. Whether they hold on the mixed-format collections common in practice, where code, markdown, tables, scientific PDFs, and prose are interleaved within one corpus, has not been measured. To study this directly, we build HetDocQA, a heterogeneous benchmark with chunker-agnostic span-overlap relevance labels and collection-disjoint splits, and pair it with MuSiQue and QASPER as homogeneous controls. We evaluate eight methods on a shared backbone, with bootstrap confidence intervals and multiple-comparison correction. A strong cross-encoder reranker accounts for most of the pipeline's quality; beyond it, only two methods yield reliable gains: query expansion and SSCC. SSCC, a per-source calibrated corrector introduced here, sets a separate acceptance threshold for each score source and helps only on heterogeneous data. The remaining reranking and pool-expansion methods in common use, among them hierarchical summarization, graph expansion, routing, and rank fusion, give no reliable gain once that reranker is present.

  • 3 authors
·
Jun 13

PRISM: Festina Lente Proactivity -- Risk-Sensitive, Uncertainty-Aware Deliberation for Proactive Agents

Proactive agents must decide not only what to say but also whether and when to intervene. Many current systems rely on brittle heuristics or indiscriminate long reasoning, which offers little control over the benefit-burden tradeoff. We formulate the problem as cost-sensitive selective intervention and present PRISM, a novel framework that couples a decision-theoretic gate with a dual-process reasoning architecture. At inference time, the agent intervenes only when a calibrated probability of user acceptance exceeds a threshold derived from asymmetric costs of missed help and false alarms. Inspired by festina lente (Latin: "make haste slowly"), we gate by an acceptance-calibrated, cost-derived threshold and invoke a resource-intensive Slow mode with counterfactual checks only near the decision boundary, concentrating computation on ambiguous and high-stakes cases. Training uses gate-aligned, schema-locked distillation: a teacher running the full PRISM pipeline provides dense, executable supervision on unlabeled interaction traces, while the student learns a response policy that is explicitly decoupled from the intervention gate to enable tunable and auditable control. On ProactiveBench, PRISM reduces false alarms by 22.78% and improves F1 by 20.14% over strong baselines. These results show that principled decision-theoretic gating, paired with selective slow reasoning and aligned distillation, yields proactive agents that are precise, computationally efficient, and controllable. To facilitate reproducibility, we release our code, models, and resources at https://prism-festinalente.github.io/; all experiments use the open-source ProactiveBench benchmark.

  • 5 authors
·
Feb 1

Per-Query Gating of LLM Rerankers for Multi-Hop Retrieval

LLM rerankers add of the order of \$0.2-0.3 per 1,000 queries and about a second of tail latency on top of a graph-augmented dense pipeline such as HippoRAG2, and on three multi-hop benchmarks they improve final-hop top-K coverage on seven of nine (dataset, K) cells, by up to +34.8 pp. We ask whether a learned per-query gate can skip the reranker where it will not help, using only features available before the LLM call (27 score and lexical statistics of the two retrieval lists plus a PCA of a small query embedding) with an executable fallback. Every choice, including the fallback and the threshold, is made inside the training fold and applied once to held-out queries, and harmful skips (the rerank would have found the target, the fallback did not) are reported next to the aggregate coverage. Across nine cells on 2WikiMultiHopQA, MuSiQue and HotpotQA the gate skips 51% of calls at an average held-out LastHop@K cost of 1.2 pp; four cells meet a pre-registered 1 pp rule, harmful skips occur in eight (190 harmful against 136 beneficial), and a random gate at the same skip rate loses 2 to 11 pp on the high-lift cells. A second rule sets each cell's threshold from a pre-specified budget on the expected harmful-skip rate over Platt-calibrated harm probabilities (ECE 0.025 after calibration, 0.094 before): at a 1 pp budget the gate skips 42% at -0.8 pp with 66 harmful skips and six cells within 1 pp, but realised harm exceeds the promise in six cells (mean 1.45 vs 0.83 pp), a selection optimism we quantify; a 0.5 pp budget realises about 1 pp. The harm probabilities are calibrated but barely discriminative (AUC 0.16 to 0.70). An earlier version reported 73% "lossless" savings; that figure rested on an oracle fallback and a wrong MuSiQue target, and we document both.

  • 1 authors
·
Sep 18

Probabilistic Assessment of Engineered Timber Reusability after Moisture Exposure

Engineered timber is pivotal to low-carbon construction, but moisture uptake during its service life can compromise structural reliability and impede reuse within a circular economy model. Despite growing interest, quantitative standards for classifying the reusability of moisture-exposed timber are still lacking. This study develops a probabilistic framework to determine the post-exposure reusability of engineered timber. Laminated specimens were soaked to full saturation, dried to 25% moisture content, and subjected to destructive three-point flexural testing. Structural integrity was quantified by a residual-performance metric that assigns 80% weight to the retained flexural modulus and 20% to the retained maximum load, benchmarked against unexposed controls. A hierarchical Bayesian multinomial logistic model with horseshoe priors, calibrated through Markov-Chain Monte-Carlo sampling, jointly infers the decision threshold separating three Modern Methods of Construction (MMC) reuse levels and predicts those levels from five field-measurable features: density, moisture content, specimen size, grain orientation, and surface hardness. Results indicate that a single wet-dry cycle preserves 70% of specimens above the 0.90 residual-performance threshold (Level 1), whereas repeated cycling lowers the mean residual to 0.78 and reallocates many specimens to Levels 2-3. The proposed framework yields quantified decision boundaries and a streamlined on-site testing protocol, providing a foundation for robust quality assurance standards.

  • 5 authors
·
May 29, 2025

YOLO-PEFT: Parameter-Efficient Fine-Tuning on YOLO Family

Generic parameter-efficient fine-tuning (PEFT) methods transferred from language models can fail silently on real-time detectors, whose heterogeneous operators and detection-specific components impose placement constraints absent from regular Transformer stacks. We propose YOLO-PEFT, a structure-aware framework that formulates adapter placement as an auditable constraint-planning problem. Given a detector graph, a PEFT request, and a resource budget, YOLO-PEFT assigns operator and semantic roles, evaluates explicit operator-validity, detector-semantic, graph-interface, and deployment predicates, records a reason code for each excluded module, and either emits a budgeted target-module plan or returns Refuse before training. Under the official VOC07+12 trainval-to-VOC07 test protocol, planner-selected RS-LoRA reaches 0.7138 and 0.7307 mAP50-95 on YOLO11s and YOLO12s, respectively, compared with 0.6428 and 0.6662 for Full-SFT. On RT-DETR-L, all seven evaluated LoRA-family configurations cross the predefined catastrophic threshold, supporting a calibrated Refuse-to-Full-SFT decision within the evaluated coverage. A controlled YOLO11 audit further shows that LoRA reduces peak training memory by 43.9 percent, although training takes 1.72 times longer. Within the evaluated detector families, placement policies, and calibration coverage, YOLO-PEFT replaces manual target-module trial and error with explicit, inspectable planning while preserving verified train-save-merge-export paths; refusal on unseen detector architectures remains an open validation problem. Project Page: github.com/Tencent/YOLO-Master

tencent Tencent
·
Aug 6 2

Unchecked or Falsely Accused? Evaluating AI-Generated Text Detectors on Kazakh

AI-text detectors are already applied to Kazakh writing, yet no study reports how often publicly available detectors flag Kazakh text whose human authorship is dated before ChatGPT. We present AINA-KK: 3,811 human passages from encyclopedic, news, web and educational pages that provably existed before ChatGPT, and 5,597 machine and gray-zone passages from 9 generators, including the Kazakh-built KazLLM and Sherkala, with script and length stress tests. Five English-trained detectors from the Hugging Face Hub flag up to 81.4% of human passages and 79.6% of published student essays; recalibrated, they catch at most 1.3% of machine text. Calibrated zero-shot detectors catch at most 50.3% of machine text from in-distribution generators. Fine-tuned encoders rank text almost perfectly, but mDeBERTa at its default threshold flags 24.4% of 50-word human excerpts, partly because machine outputs are shorter than human text; calibrated, it catches only 47.1% of KazLLM and 62.7% of Sherkala passages. Without script handling, machine-transliterated Latin-2021 text defeats every detector. Fixes that need no new data (calibration, script canonicalization, script-augmented and length-matched training, and an OR with a zero-shot scorer) form AINA-KK-Det. On test it flags 0.9% of human passages (95% CI 0.4-2.7; 1.5% on development data) and catches 85.2% of machine text from known generators, 92.9% from two unseen Gemini-family Gemma-4 models and 61.3% from the Kazakh-built ones; its configuration was chosen with test results in view. We release the data, code and detector at https://github.com/ozxmn/aina-kk.

  • 1 authors
·
Oct 11

Seq-UPS: Sequential Uncertainty-aware Pseudo-label Selection for Semi-Supervised Text Recognition

This paper looks at semi-supervised learning (SSL) for image-based text recognition. One of the most popular SSL approaches is pseudo-labeling (PL). PL approaches assign labels to unlabeled data before re-training the model with a combination of labeled and pseudo-labeled data. However, PL methods are severely degraded by noise and are prone to over-fitting to noisy labels, due to the inclusion of erroneous high confidence pseudo-labels generated from poorly calibrated models, thus, rendering threshold-based selection ineffective. Moreover, the combinatorial complexity of the hypothesis space and the error accumulation due to multiple incorrect autoregressive steps posit pseudo-labeling challenging for sequence models. To this end, we propose a pseudo-label generation and an uncertainty-based data selection framework for semi-supervised text recognition. We first use Beam-Search inference to yield highly probable hypotheses to assign pseudo-labels to the unlabelled examples. Then we adopt an ensemble of models, sampled by applying dropout, to obtain a robust estimate of the uncertainty associated with the prediction, considering both the character-level and word-level predictive distribution to select good quality pseudo-labels. Extensive experiments on several benchmark handwriting and scene-text datasets show that our method outperforms the baseline approaches and the previous state-of-the-art semi-supervised text-recognition methods.

Purdue Purdue University
·
Oct 5, 2022

Look Before you Leap: Estimating LLM Benchmark Scores from Descriptions

Progress in large language models is constrained by an evaluation bottleneck: build a benchmark, evaluate models and settings, then iterate. We therefore ask a simple question: can we forecast outcomes before running any experiments? We study text-only performance forecasting: estimating a model's score from a redacted task description and intended configuration, with no access to dataset instances. To support systematic study, we curate PRECOG, a corpus of redacted description-performance pairs spanning diverse tasks, domains, and metrics. Experiments show the task is challenging but feasible: models equipped with a retrieval module that excludes source papers achieve moderate prediction performance with well-calibrated uncertainty, reaching mean absolute error as low as 8.7 on the Accuracy subset at high-confidence thresholds. Our analysis indicates that stronger reasoning models engage in diverse, iterative querying, whereas current open-source models lag and often skip retrieval or gather evidence with limited diversity. We further test a zero-leakage setting, forecasting on newly released datasets or experiments before their papers are indexed, where GPT-5 with built-in web search still attains nontrivial prediction accuracy. Overall, our corpus and analyses offer an initial step toward open-ended anticipatory evaluation, supporting difficulty estimation and smarter experiment prioritization.

  • 4 authors
·
Sep 24, 2025

FlexGuard: Continuous Risk Scoring for Strictness-Adaptive LLM Content Moderation

Ensuring the safety of LLM-generated content is essential for real-world deployment. Most existing guardrail models formulate moderation as a fixed binary classification task, implicitly assuming a fixed definition of harmfulness. In practice, enforcement strictness - how conservatively harmfulness is defined and enforced - varies across platforms and evolves over time, making binary moderators brittle under shifting requirements. We first introduce FlexBench, a strictness-adaptive LLM moderation benchmark that enables controlled evaluation under multiple strictness regimes. Experiments on FlexBench reveal substantial cross-strictness inconsistency in existing moderators: models that perform well under one regime can degrade substantially under others, limiting their practical usability. To address this, we propose FlexGuard, an LLM-based moderator that outputs a calibrated continuous risk score reflecting risk severity and supports strictness-specific decisions via thresholding. We train FlexGuard via risk-alignment optimization to improve score-severity consistency and provide practical threshold selection strategies to adapt to target strictness at deployment. Experiments on FlexBench and public benchmarks demonstrate that FlexGuard achieves higher moderation accuracy and substantially improved robustness under varying strictness. We release the source code and data to support reproducibility.

  • 4 authors
·
Feb 26

Low-Resource Safety Failures Are Action Failures, Not Representation Failures

Language models often answer harmful requests in low-resource languages (LRLs) that they refuse in high-resource languages (HRLs). Across three instruction-tuned models and 23 languages, harmful refusal falls from 87.9% in HRLs to 43.9% in LRLs, while harmless refusal remains low. A common explanation is that models represent harmfulness weakly in LRLs. We test whether harmfulness is instead represented but does not reliably produce refusal. Across three models, a harmfulness direction learned from HRL activations still separates harmful from harmless LRL prompts, showing that complete absence of harmfulness information cannot explain many failures. However, harmfulness scores shift downward for LRL prompts, making harmful prompts less likely to reach the range associated with refusal. Motivated by this shift, we train a low-rank logistic classifier on HRL activations and calibrate its threshold with a few target-language examples. During generation, the classifier conditionally adds or ablates the HRL harmfulness direction. With the same HRL data and 32 target-language examples per class, CAST remains limited by low harmful refusal and AdaSteer by high harmless refusal, yielding mean refusal selectivity (Δ = harmful - harmless refusal) of 33.6 and 6.8, respectively. Our intervention reaches 54.5 while preserving MMLU utility. HRL-only calibration improves selectivity for Qwen and Gemma, whereas Llama benefits from target-language calibration. These results show that recalibrating existing representations can offer a training-free method for repairing low-resource safety failures.

  • 3 authors
·
Oct 4

CRC-Router: Risk-Constrained Routing for Medical Agentic AI Systems

Agentic AI systems are increasingly being explored in medical imaging to improve throughput and reduce clinician workload; however, safe deployment remains challenging because autonomous errors may propagate into downstream clinical decisions. A central requirement is therefore not only strong predictive performance, but also a reliable routing mechanism that determines when the system should proceed autonomously and when a case should be escalated for further review. To address this gap, we propose CRC-Router, a risk-constrained, uncertainty-aware routing module that is applicable to both conventional medical prediction models and agentic medical AI systems. CRC-Router combines multiple complementary uncertainty signals with the predictive score to construct a per-finding routing feature vector, maps this vector to an estimated wrong-accept risk using a lightweight per-finding risk model, and then applies Conformal Risk Control (CRC) to calibrate acceptance thresholds under a user-specified risk target. Instantiated on chest X-ray multi-finding triage using the NIH ChestX-ray14 dataset, CRC-Router achieves the strongest empirical risk--coverage trade-off among the evaluated baselines, both as a standalone routing layer and as a plug-in module integrated with the state-of-the-art MedRAX agent. These results demonstrate both the effectiveness of CRC-Router in selective medical automation and its modular, model-agnostic compatibility with existing predictive and agentic medical pipelines. Code is publicly available at https://github.com/XLIAaron/CRC-Router

  • 8 authors
·
Sep 24

Measuring Calibration in Deep Learning

Overconfidence and underconfidence in machine learning classifiers is measured by calibration: the degree to which the probabilities predicted for each class match the accuracy of the classifier on that prediction. How one measures calibration remains a challenge: expected calibration error, the most popular metric, has numerous flaws which we outline, and there is no clear empirical understanding of how its choices affect conclusions in practice, and what recommendations there are to counteract its flaws. In this paper, we perform a comprehensive empirical study of choices in calibration measures including measuring all probabilities rather than just the maximum prediction, thresholding probability values, class conditionality, number of bins, bins that are adaptive to the datapoint density, and the norm used to compare accuracies to confidences. To analyze the sensitivity of calibration measures, we study the impact of optimizing directly for each variant with recalibration techniques. Across MNIST, Fashion MNIST, CIFAR-10/100, and ImageNet, we find that conclusions on the rank ordering of recalibration methods is drastically impacted by the choice of calibration measure. We find that conditioning on the class leads to more effective calibration evaluations, and that using the L2 norm rather than the L1 norm improves both optimization for calibration metrics and the rank correlation measuring metric consistency. Adaptive binning schemes lead to more stablity of metric rank ordering when the number of bins vary, and is also recommended. We open source a library for the use of our calibration measures.

  • 7 authors
·
Aug 6, 2020

SCOPE: Selective Conformal Optimized Pairwise LLM Judging

Large language models (LLMs) are increasingly used as judges to replace costly human preference labels in pairwise evaluation. Despite their practicality, LLM judges remain prone to miscalibration and systematic biases. This paper proposes SCOPE (Selective Conformal Optimized Pairwise Evaluation), a framework for selective pairwise judging with finite-sample statistical guarantees. Under exchangeability, SCOPE calibrates an acceptance threshold such that the error rate among non-abstained judgments is at most a user-specified level α. To provide SCOPE with a bias-neutral uncertainty signal, we introduce Bidirectional Preference Entropy (BPE), which queries the judge under both response positions, aggregates the implied preference probabilities to enforce invariance to response order, and converts the aggregated probability into an entropy-based uncertainty score. Across MT-Bench, RewardBench, and Chatbot Arena, BPE improves uncertainty quality over standard confidence proxies, providing a stronger selection signal that enables SCOPE to consistently meet the target risk level while retaining good coverage across judge scales. In particular, at α= 0.10, SCOPE consistently satisfies the risk bound across all benchmarks and judge scales (empirical risk approx 0.097 to 0.099), while retaining substantial coverage, reaching 0.89 on RewardBench with Qwen-14B and 0.98 on RewardBench with Qwen-32B. Compared to naïve baselines, SCOPE accepts up to 2.4times more judgments on MT-Bench with Qwen-7B under the same target risk constraint, demonstrating that BPE enables reliable and high-coverage LLM-based evaluation.

  • 3 authors
·
Feb 18

Inference-Time Conformal Reasoning with Valid Factuality Control for Large Language Models

Large language models (LLMs) increasingly perform multi-step reasoning, where intermediate claims form implicit directed acyclic graphs whose node correctness is structurally conditioned on their ancestors. This makes factuality uncertainty structural, rather than a trivial accumulation of node-wise errors, and necessitates inference-time uncertainty quantification over the reasoning structure. While conformal prediction (CP) offers flexible user-specified factuality control, existing work remains post-hoc and cannot intervene during generation. To fill the gap between CP's flexibility and its post-hoc limitation, we propose an Inference-Time Conformal Reasoning (ITCR) framework that integrates CP directly into reasoning graph generation. ITCR learns a structure-level factuality uncertainty function that aggregates claim-level factuality signals over reasoning graphs without complex modeling assumptions. We then design the non-conformity score based on graph-level factuality uncertainty and calibrate the conformal threshold to decide when to stop generation. We theoretically show such generation is nested, yielding valid coverage guarantees for factuality control. Experiments over multiple datasets and coverage objectives demonstrate empirically valid coverage. In downstream reasoning tasks, inference-time calibrated graphs yield more accurate generation than post-hoc pruned graphs.

  • 4 authors
·
Jun 27

When PTQ Calibration Helps or Hurts: Dose-Driven Collapse and Validation-Gated Abstention in Quantized Text Embedders

Learned post-training quantization (PTQ) calibration fits quantizer parameters on a small calibration set, and for a text embedder the natural supervision is retrieval relevance. We ask two questions that are usually conflated: how far to calibrate, and whether to calibrate at all. Across three embedders (e5-base-v2, BGE-M3, EmbeddingGemma-300m) at 4-bit weight quantization, 161 calibration runs, the outcome of a fixed procedure ranges from a gain to catastrophic degradation, and the driver is the dose of the update, measured as the functional displacement D_func of the quantized model from its uncalibrated start. On EmbeddingGemma the same learning-rate grid gives +1.47 to -71.53 points of in-domain NDCG@10 change; a label-free cosine objective collapses at the same dose (-69.52 vs -71.63), so the collapse is a property of learnable calibration rather than of retrieval supervision. Objective sensitivity is heterogeneous: on e5, ranking calibration is substantially more robust than cosine at matched displacement (+14.16 +/- 4.34 pp, 95% t-CI [+3.37, +24.96], three paired slates), whereas on BGE-M3 the ordering is directionally opposite but unresolved ([-5.18, +2.38]). Dose is the main axis but not a sufficient statistic: equal D_func reached along different paths differs by up to 1.61 pp. We therefore treat calibration as an optional intervention. A pre-registered rule that accepts a checkpoint only when a per-query paired confidence interval on a 25% calibration-validation split clears the uncalibrated baseline was run on 6 independent slates and five arms. It never selected a checkpoint below its slate's baseline (0 harmful selections in 30 decisions), retained 8 of 11 beneficial updates, and paid a mean oracle regret of 1.52 pp with one abstention that forfeited 11.13 pp, a safety-opportunity trade-off, not a free improvement. We report what remains open, including label efficiency.

  • 4 authors
·
Oct 18

SimplexUQ: An Evaluation Framework and Benchmark for Conformal Uncertainty on Simplex-Valued Predictions

Conformal prediction guarantees marginal coverage, but a single calibration threshold can still spread that coverage unevenly, over-covering easy regions and under-covering hard ones. SimplexUQ is, to our knowledge, the first benchmark and reproducible protocol for measuring this allocation problem on simplex-valued predictions; it compares existing conformal wrappers rather than proposing a new one. Its task suite, SimplexTasks-12, combines six controlled synthetic regimes with six frozen-predictor real tasks spanning class probabilities, topic mixtures, spectral abundances, cell-type fractions, age distributions, and emotion mixtures. Each comparison fixes the predictor, score, and response-free stratification map, varies only the wrapper, and reports marginal coverage, worst-stratum coverage, max disparity, and within-task radius and compute. Global calibration can look valid while failing badly: on CIFAR-10 it attains 0.900 marginal coverage but only 0.542 in the worst entropy stratum, and Mondrian calibration raises that stratum to 0.886 while reducing max disparity from 0.358 to 0.022. No wrapper dominates, however. Under smooth synthetic heterogeneity, several repairs are competitive; fixed-map analyses show that rankings depend on the evaluation groups and protocol; and in a 12-task comparison, Mondrian has lower disparity on its single target partition for all 12 tasks, whereas BatchMVP has lower disparity over overlapping groups on five. These are empirical comparisons, not new coverage guarantees. A controlled predictor-bias sweep shows that removing predictor bias only partly reduces global-threshold disparity. We release task cards, result provenance, permitted derived arrays, and rebuild instructions, and treat wrapper selection as a diagnostic comparison rather than a universal ranking.

  • 3 authors
·
Sep 29

h-calibration: Rethinking Classifier Recalibration with Probabilistic Error-Bounded Objective

Deep neural networks have demonstrated remarkable performance across numerous learning tasks but often suffer from miscalibration, resulting in unreliable probability outputs. This has inspired many recent works on mitigating miscalibration, particularly through post-hoc recalibration methods that aim to obtain calibrated probabilities without sacrificing the classification performance of pre-trained models. In this study, we summarize and categorize previous works into three general strategies: intuitively designed methods, binning-based methods, and methods based on formulations of ideal calibration. Through theoretical and practical analysis, we highlight ten common limitations in previous approaches. To address these limitations, we propose a probabilistic learning framework for calibration called h-calibration, which theoretically constructs an equivalent learning formulation for canonical calibration with boundedness. On this basis, we design a simple yet effective post-hoc calibration algorithm. Our method not only overcomes the ten identified limitations but also achieves markedly better performance than traditional methods, as validated by extensive experiments. We further analyze, both theoretically and experimentally, the relationship and advantages of our learning objective compared to traditional proper scoring rule. In summary, our probabilistic framework derives an approximately equivalent differentiable objective for learning error-bounded calibrated probabilities, elucidating the correspondence and convergence properties of computational statistics with respect to theoretical bounds in canonical calibration. The theoretical effectiveness is verified on standard post-hoc calibration benchmarks by achieving state-of-the-art performance. This research offers valuable reference for learning reliable likelihood in related fields.

  • 6 authors
·
Jun 22, 2025

Verified Uncertainty Calibration

Applications such as weather forecasting and personalized medicine demand models that output calibrated probability estimates---those representative of the true likelihood of a prediction. Most models are not calibrated out of the box but are recalibrated by post-processing model outputs. We find in this work that popular recalibration methods like Platt scaling and temperature scaling are (i) less calibrated than reported, and (ii) current techniques cannot estimate how miscalibrated they are. An alternative method, histogram binning, has measurable calibration error but is sample inefficient---it requires O(B/ε^2) samples, compared to O(1/ε^2) for scaling methods, where B is the number of distinct probabilities the model can output. To get the best of both worlds, we introduce the scaling-binning calibrator, which first fits a parametric function to reduce variance and then bins the function values to actually ensure calibration. This requires only O(1/ε^2 + B) samples. Next, we show that we can estimate a model's calibration error more accurately using an estimator from the meteorological community---or equivalently measure its calibration error with fewer samples (O(B) instead of O(B)). We validate our approach with multiclass calibration experiments on CIFAR-10 and ImageNet, where we obtain a 35% lower calibration error than histogram binning and, unlike scaling methods, guarantees on true calibration. In these experiments, we also estimate the calibration error and ECE more accurately than the commonly used plugin estimators. We implement all these methods in a Python library: https://pypi.org/project/uncertainty-calibration

  • 3 authors
·
Sep 23, 2019

On Calibration of Object Detectors: Pitfalls, Evaluation and Baselines

Reliable usage of object detectors require them to be calibrated -- a crucial problem that requires careful attention. Recent approaches towards this involve (1) designing new loss functions to obtain calibrated detectors by training them from scratch, and (2) post-hoc Temperature Scaling (TS) that learns to scale the likelihood of a trained detector to output calibrated predictions. These approaches are then evaluated based on a combination of Detection Expected Calibration Error (D-ECE) and Average Precision. In this work, via extensive analysis and insights, we highlight that these recent evaluation frameworks, evaluation metrics, and the use of TS have notable drawbacks leading to incorrect conclusions. As a step towards fixing these issues, we propose a principled evaluation framework to jointly measure calibration and accuracy of object detectors. We also tailor efficient and easy-to-use post-hoc calibration approaches such as Platt Scaling and Isotonic Regression specifically for object detection task. Contrary to the common notion, our experiments show that once designed and evaluated properly, post-hoc calibrators, which are extremely cheap to build and use, are much more powerful and effective than the recent train-time calibration methods. To illustrate, D-DETR with our post-hoc Isotonic Regression calibrator outperforms the recent train-time state-of-the-art calibration method Cal-DETR by more than 7 D-ECE on the COCO dataset. Additionally, we propose improved versions of the recently proposed Localization-aware ECE and show the efficacy of our method on these metrics as well. Code is available at: https://github.com/fiveai/detection_calibration.

  • 4 authors
·
May 30, 2024

Learning to Detect Shifts in Unstructured Data Streams via Monitoring-by-Betting

Continuous monitoring of generative AI systems requires detecting distribution shifts in the unstructured data they ingest and generate. Existing monitoring methods typically control false alarms by requiring the discrepancy measure used to detect shifts to be *pre-specified*. However, for unstructured data, pre-specification creates a sensitivity--breadth trade-off: a narrowly defined measure is sensitive to the anticipated shift but may miss unexpected changes, whereas a broadly defined measure may detect a wider range of shifts but with reduced sensitivity. Although adaptively hypothesizing the shift direction can overcome this limitation, such data-dependent adaptation generally invalidates conventional false-alarm guarantees. We reconcile adaptivity with statistical validity through testing-by-betting, reframing sequential changepoint detection as a betting game in which the bettor continually updates its hypothesized shift direction, or witness function, as new observations arrive. We introduce *asymptotic e-detectors*, which approximate the pre-change distribution using a calibration set of size m, and prove that they asymptotically control the average run length to a false alarm as m increases. Empirically, calibration sets as small as m=30 provide effective control. We then develop adaptive monitoring strategies ranging from black-box bettors that learn from embeddings to interpretable bettors based on LLM-proposed concepts. Across three real-world datasets, including data from a clinical AI scribe deployment, adaptive monitors detect shifts faster than fixed monitors at the same false-alarm level. Further, when the true shifts are sparse and interpretable, LLM-concept monitors provide the fastest and most interpretable detection, while also guiding root-cause diagnosis for complex shifts.

  • 11 authors
·
Oct 11

Adaptive Alarm Threshold Prediction in 4G Mobile Networks: A Percentile-Guided Deep Learning Framework with Interpretable Outputs

In mobile telecommunications, alarms act as early warning signals. They are triggered when a cell, the basic unit of radio coverage, shuts down or behaves abnormally. This signals a degradation in service quality, which directly affects the customer experience. To fix the issue, operators rely on preset thresholds to decide when an engineer should be sent out. In practice, these thresholds are set manually and remain fixed regardless of the time of day, traffic levels, or overall network conditions. This often leads to serious faults slipping through during busy hours, while minor issues can cause unnecessary callouts when the network is quiet. This paper presents a machine learning framework that automatically predicts four alarm thresholds, audit window duration, inactive time limit, total fluctuation count, and per hour fluctuation limit, from live network behavior. Since no ground truth labels exist for thresholds, we introduce a percentile guided label derivation strategy and evaluate four models on an anonymized dataset of 10,648 cells across three vendors and nine regions from a real 4G network, comprising a Gradient Boosted Trees baseline, a CNN-BiLSTM with attention, the proposed PCTN, and an iTransformer. PCTN performs the best overall with respect to three of the four targets, outperforming a state-of-the-art iTransformer while using 83 percent fewer parameters. Its mixed output heads and dynamic alpha mechanism produce thresholds that are both accurate and interpretable, allowing operators to inspect and adjust the learned policy without retraining. All comparisons are statistically significant at p < 0.001. The framework undergoes daily retraining using new data, which enables the thresholds to constantly adjust to changes in the network.

  • 3 authors
·
Apr 3

CalArena: A Large-Scale Post-Hoc Calibration Benchmark

Reliable probability estimates are critical in many machine learning applications, yet modern classifiers are often poorly calibrated. Post-hoc calibration provides a simple and widely used solution, but the large number of proposed methods, combined with small-scale and inconsistent evaluations, makes it difficult to determine which approaches are truly effective in practice. We introduce a large-scale, standardized benchmark for post-hoc calibration, covering nearly 2000 experiments across tabular and computer vision tasks, including binary, multiclass, and large-scale classification settings. Our benchmark aggregates predictions from a diverse set of classical models, modern deep learning architectures, and foundation models, and provides unified, reproducible implementations of dozens of calibration methods within a common evaluation framework. We argue that Post-Hoc Improvement (PHI) in proper scoring rules offers a principled alternative to traditional calibration error estimators for comparing post-hoc methods, capturing both calibration quality and potential degradation to the model's predictive performance. Using this framework, we conduct the most comprehensive empirical study of post-hoc calibration to date. Our results reveal consistent patterns across domains: smooth calibration functions outperform binning-based approaches, dedicated multiclass methods are essential in high-dimensional settings, and generic machine learning models are not competitive without calibration-specific design. To facilitate future research, we release all data, code, and evaluation tools, providing a plug-and-play benchmark for developing and comparing calibration methods.

  • 4 authors
·
May 27

When Does Risk-Controlled Throughput Forecasting Actually Control Risk? A Reproduction Study on LEO Satellite Link

Risk-budgeted throughput forecasting converts a point forecast of LEO satellite capacity into a conservative bound whose overestimation rate is held at a user-specified budget, so that admission control can act on it safely. We reproduce the state of the art, BG-CFQS, on the three Starlink traces its authors used and on an independent fourth dataset, and report four findings. First, its risk guarantee is reproducible but conditional on the calibration split: under a random split we recover every published quantity, including risk-pass on 3 of 3 datasets, but under a contiguous temporal split, which is what a deployed terminal faces, it holds on 1 of 3. Second, independent of any split, the method cannot serve a risk budget below 0.15, the low end of its candidate quantile set. On all four datasets the achieved rate is identical at budgets of 0.05, 0.10 and 0.15, overshooting by 1.4x to 3.7x at the tightest. Third, static conformal calibration cannot improve a ranking, only shift an operating point; we confirm this exactly on latency spike detection, where online calibration improves AUPRC by up to 51%. Fourth, group-conditional online recalibration splits into two mechanisms. The online part replicates: it tracks the budget within 1% from 0.05 to 0.35, where static calibration is 16 to 21% off. The group-conditional part does not: on three of four datasets, conditioning on any available covariate, including measured serving-satellite geometry, is worse than not conditioning, and our attempt to predict which datasets benefit failed. What is new here is empirical, not algorithmic: we claim the findings, not the methods.

  • 5 authors
·
Nov 3

TypedBench: A Benchmark for Calibration, Framing Sensitivity, and Cost in System One Decision Models

System One models output calibrated probabilities over typed answers such as categorical choices, ordinal levels, or binary outcomes, via a non-generative interface. Software can act on these probabilities through thresholds, cost-weighted choices, and escalation rules. Consequently, if these probabilities are miscalibrated or wording-sensitive, the software ma take unintended actions leaving human operators with no textual rationale to inspect. Current evaluations largely report accuracy and calibration on public classification datasets without a clear reference. We present TypedBench, a benchmark for typed decision models like Jev built from seven policy-labelled generators and nine evaluation suites. We report accuracy as median and range across paraphrases, and calibration error relative to the finite-sample noise floor of a matched, perfectly calibrated predictor. We assess probability quality through selective prediction, ordinal proper scoring rules, and realised cost under asymmetric cost matrices. We evaluate a hosted model, an open encoder, and a family of open decoders spanning 0.8B-9B parameters on identical items. The hosted model follows the stated policy but is wording-sensitive and systematically underconfident; under asymmetric costs, using its probabilities can be worse than taking its top answer. Decoders route exactly and are slow as options or questions are added. The decoder is least accurate on policy questions and degrades with more options. Overall, typed decision models must be evaluated jointly on policy adherence, wording robustness, probability quality, and induced decision outcomes.

  • 4 authors
·
Oct 8

Conformal Segmentation in Industrial Surface Defect Detection with Statistical Guarantees

In industrial settings, surface defects on steel can significantly compromise its service life and elevate potential safety risks. Traditional defect detection methods predominantly rely on manual inspection, which suffers from low efficiency and high costs. Although automated defect detection approaches based on Convolutional Neural Networks(e.g., Mask R-CNN) have advanced rapidly, their reliability remains challenged due to data annotation uncertainties during deep model training and overfitting issues. These limitations may lead to detection deviations when processing the given new test samples, rendering automated detection processes unreliable. To address this challenge, we first evaluate the detection model's practical performance through calibration data that satisfies the independent and identically distributed (i.i.d) condition with test data. Specifically, we define a loss function for each calibration sample to quantify detection error rates, such as the complement of recall rate and false discovery rate. Subsequently, we derive a statistically rigorous threshold based on a user-defined risk level to identify high-probability defective pixels in test images, thereby constructing prediction sets (e.g., defect regions). This methodology ensures that the expected error rate (mean error rate) on the test set remains strictly bounced by the predefined risk level. Additionally, we observe a negative correlation between the average prediction set size and the risk level on the test set, establishing a statistically rigorous metric for assessing detection model uncertainty. Furthermore, our study demonstrates robust and efficient control over the expected test set error rate across varying calibration-to-test partitioning ratios, validating the method's adaptability and operational effectiveness.

  • 2 authors
·
Apr 23, 2025

Optimizing Calibration by Gaining Aware of Prediction Correctness

Model calibration aims to align confidence with prediction correctness. The Cross-Entropy (CE) loss is widely used for calibrator training, which enforces the model to increase confidence on the ground truth class. However, we find the CE loss has intrinsic limitations. For example, for a narrow misclassification, a calibrator trained by the CE loss often produces high confidence on the wrongly predicted class (e.g., a test sample is wrongly classified and its softmax score on the ground truth class is around 0.4), which is undesirable. In this paper, we propose a new post-hoc calibration objective derived from the aim of calibration. Intuitively, the proposed objective function asks that the calibrator decrease model confidence on wrongly predicted samples and increase confidence on correctly predicted samples. Because a sample itself has insufficient ability to indicate correctness, we use its transformed versions (e.g., rotated, greyscaled and color-jittered) during calibrator training. Trained on an in-distribution validation set and tested with isolated, individual test samples, our method achieves competitive calibration performance on both in-distribution and out-of-distribution test sets compared with the state of the art. Further, our analysis points out the difference between our method and commonly used objectives such as CE loss and mean square error loss, where the latters sometimes deviates from the calibration aim.

  • 5 authors
·
Apr 19, 2024

Using Machine Learning to Calibrate Storm-Scale Probabilistic Guidance of Severe Weather Hazards in the Warn-on-Forecast System

A primary goal of the National Oceanic and Atmospheric Administration (NOAA) Warn-on-Forecast (WoF) project is to provide rapidly updating probabilistic guidance to human forecasters for short-term (e.g., 0-3 h) severe weather forecasts. Maximizing the usefulness of probabilistic severe weather guidance from an ensemble of convection-allowing model forecasts requires calibration. In this study, we compare the skill of a simple method using updraft helicity against a series of machine learning (ML) algorithms for calibrating WoFS severe weather guidance. ML models are often used to calibrate severe weather guidance since they leverage multiple variables and discover useful patterns in complex datasets. \indent Our dataset includes WoF System (WoFS) ensemble forecasts available every 5 minutes out to 150 min of lead time from the 2017-2019 NOAA Hazardous Weather Testbed Spring Forecasting Experiments (81 dates). Using a novel ensemble storm track identification method, we extracted three sets of predictors from the WoFS forecasts: intra-storm state variables, near-storm environment variables, and morphological attributes of the ensemble storm tracks. We then trained random forests, gradient-boosted trees, and logistic regression algorithms to predict which WoFS 30-min ensemble storm tracks will correspond to a tornado, severe hail, and/or severe wind report. For the simple method, we extracted the ensemble probability of 2-5 km updraft helicity (UH) exceeding a threshold (tuned per severe weather hazard) from each ensemble storm track. The three ML algorithms discriminated well for all three hazards and produced more reliable probabilities than the UH-based predictions. Overall, the results suggest that ML-based calibrations of dynamical ensemble output can improve short term, storm-scale severe weather probabilistic guidance

  • 5 authors
·
Nov 11, 2020

YOCO: You Only Calibrate Once for Accurate Extrinsic Parameter in LiDAR-Camera Systems

In a multi-sensor fusion system composed of cameras and LiDAR, precise extrinsic calibration contributes to the system's long-term stability and accurate perception of the environment. However, methods based on extracting and registering corresponding points still face challenges in terms of automation and precision. This paper proposes a novel fully automatic extrinsic calibration method for LiDAR-camera systems that circumvents the need for corresponding point registration. In our approach, a novel algorithm to extract required LiDAR correspondence point is proposed. This method can effectively filter out irrelevant points by computing the orientation of plane point clouds and extracting points by applying distance- and density-based thresholds. We avoid the need for corresponding point registration by introducing extrinsic parameters between the LiDAR and camera into the projection of extracted points and constructing co-planar constraints. These parameters are then optimized to solve for the extrinsic. We validated our method across multiple sets of LiDAR-camera systems. In synthetic experiments, our method demonstrates superior performance compared to current calibration techniques. Real-world data experiments further confirm the precision and robustness of the proposed algorithm, with average rotation and translation calibration errors between LiDAR and camera of less than 0.05 degree and 0.015m, respectively. This method enables automatic and accurate extrinsic calibration in a single one step, emphasizing the potential of calibration algorithms beyond using corresponding point registration to enhance the automation and precision of LiDAR-camera system calibration.

  • 4 authors
·
Jul 25, 2024

Optimistic Feasible Search for Closed-Loop Fair Threshold Decision-Making

Closed-loop decision-making systems (e.g., lending, screening, or recidivism risk assessment) often operate under fairness and service constraints while inducing feedback effects: decisions change who appears in the future, yielding non-stationary data and potentially amplifying disparities. We study online learning of a one-dimensional threshold policy from bandit feedback under demographic parity (DP) and, optionally, service-rate constraints. The learner observes only a scalar score each round and selects a threshold; reward and constraint residuals are revealed only for the chosen threshold. We propose Optimistic Feasible Search (OFS), a simple grid-based method that maintains confidence bounds for reward and constraint residuals for each candidate threshold. At each round, OFS selects a threshold that appears feasible under confidence bounds and, among those, maximizes optimistic reward; if no threshold appears feasible, OFS selects the threshold minimizing optimistic constraint violation. This design directly targets feasible high-utility thresholds and is particularly effective for low-dimensional, interpretable policy classes where discretization is natural. We evaluate OFS on (i) a synthetic closed-loop benchmark with stable contraction dynamics and (ii) two semi-synthetic closed-loop benchmarks grounded in German Credit and COMPAS, constructed by training a score model and feeding group-dependent acceptance decisions back into population composition. Across all environments, OFS achieves higher reward with smaller cumulative constraint violation than unconstrained and primal-dual bandit baselines, and is near-oracle relative to the best feasible fixed threshold under the same sweep procedure. Experiments are reproducible and organized with double-blind-friendly relative outputs.

  • 1 authors
·
Dec 26, 2025

Calibrated Decisions at Scale: Converting Police Crash Narratives into Probabilistic Crash Variables with a System One Model (Jev)

Crash datasets that carry an investigator narrative hold information the coded fields omit. Coding those narratives at scale has been blocked by three obstacles. Frontier large language models are costly at that scale, their generated text cannot be verified, and no rule says how much output a human must check. This paper formulates narrative coding as gated, typed decisions answered by Jev, a System One model that returns probabilities over analyst-defined options and generates no text. A screen covered 499,500 Texas narratives and 195,857 were coded with a 27-question schema. Cost is governed by schema size rather than narrative length. The probabilities are audited against coded fields and against 2,416 blinded human judgments drawn under a stated sampling design. Two frontier large language models are benchmarked on the same records. Against human labels the typed model attains an F1 of 0.908. One frontier model gains 0.059 and the other is indistinguishable from it. Calibration varies by model rather than by paradigm, so each model must be audited. Recalibration on the same labels reduces calibration error by a factor of 3.3. Agreement with coded fields understates fidelity to the narrative by a median of 0.26 in kappa. A resolution-floor bound covers any model that reports probabilities on a discrete grid. A review budget over flagged records gives the records a human must read per variable and per year. Adding the calibrated variables to the coded fields raises the injury and fatal crashes attributed to nine factors by 10,747 per year.

  • 2 authors
·
Sep 20

Learning Conformal Abstention Policies for Adaptive Risk Management in Large Language and Vision-Language Models

Large Language and Vision-Language Models (LLMs/VLMs) are increasingly used in safety-critical applications, yet their opaque decision-making complicates risk assessment and reliability. Uncertainty quantification (UQ) helps assess prediction confidence and enables abstention when uncertainty is high. Conformal prediction (CP), a leading UQ method, provides statistical guarantees but relies on static thresholds, which fail to adapt to task complexity and evolving data distributions, leading to suboptimal trade-offs in accuracy, coverage, and informativeness. To address this, we propose learnable conformal abstention, integrating reinforcement learning (RL) with CP to optimize abstention thresholds dynamically. By treating CP thresholds as adaptive actions, our approach balances multiple objectives, minimizing prediction set size while maintaining reliable coverage. Extensive evaluations across diverse LLM/VLM benchmarks show our method outperforms Least Ambiguous Classifiers (LAC) and Adaptive Prediction Sets (APS), improving accuracy by up to 3.2%, boosting AUROC for hallucination detection by 22.19%, enhancing uncertainty-guided selective generation (AUARC) by 21.17%, and reducing calibration error by 70%-85%. These improvements hold across multiple models and datasets while consistently meeting the 90% coverage target, establishing our approach as a more effective and flexible solution for reliable decision-making in safety-critical applications. The code is available at: {https://github.com/sinatayebati/vlm-uncertainty}.

  • 6 authors
·
Feb 8, 2025 2

Self-Calibration and Bilinear Inverse Problems via Linear Least Squares

Whenever we use devices to take measurements, calibration is indispensable. While the purpose of calibration is to reduce bias and uncertainty in the measurements, it can be quite difficult, expensive, and sometimes even impossible to implement. We study a challenging problem called self-calibration, i.e., the task of designing an algorithm for devices so that the algorithm is able to perform calibration automatically. More precisely, we consider the setup y = A(d) x + epsilon where only partial information about the sensing matrix A(d) is known and where A(d) linearly depends on d. The goal is to estimate the calibration parameter d (resolve the uncertainty in the sensing process) and the signal/object of interests x simultaneously. For three different models of practical relevance, we show how such a bilinear inverse problem, including blind deconvolution as an important example, can be solved via a simple linear least squares approach. As a consequence, the proposed algorithms are numerically extremely efficient, thus potentially allowing for real-time deployment. We also present a variation of the least squares approach, which leads to a~spectral method, where the solution to the bilinear inverse problem can be found by computing the singular vector associated with the smallest singular value of a certain matrix derived from the bilinear system. Explicit theoretical guarantees and stability theory are derived for both techniques; and the number of sampling complexity is nearly optimal (up to a poly-log factor). Applications in imaging sciences and signal processing are discussed and numerical simulations are presented to demonstrate the effectiveness and efficiency of our approach.

  • 2 authors
·
Nov 13, 2016

Verified, not generated: expert-verified AI study materials and the distribution of learning gains in a university course

Experimental studies of generative AI in education mostly report average effects, yet field evidence shows that AI can narrow attainment gaps or widen them. We argue that the direction depends on the judgement burden, the expertise a learner must supply to screen AI output before learning from it, and that expert verification before release moves this burden from students to an accountable tutor. We test the argument in a two-cohort difference-in-differences design in which one half of a compulsory firstyear university economics course received AI-generated podcasts, FAQs and quiz-based study guides, produced with a source-grounded model and checked by a named graduate teaching assistant (170 students; 340 examination marks). Access was associated with a 2.34-mark advantage on a 50-mark component. The share of marks below the upper-second classification boundary fell by 24.7 percentage points relative to the counterfactual, effects were significant at every threshold from 23 to 31 marks and at none above, and roughly three-quarters of the average originated in the bottom quintile. The threshold estimate is robust to removing the lowest-scoring students from the pre-intervention cohort; the average effect is not. Interviews and feedback from 36 students indicate that the verification label gave students a reason to engage with AI-generated material without ending their scrutiny of it. Evaluations of AI learning resources that report only mean effects cannot detect whether the students the resources are meant to help are the ones who gain.

  • 2 authors
·
Oct 4

Calibrated Evidence Inference With a Small, Permissively Licensed Open Model

Systematic reviewers who use a model to read randomized trials need each answer and a measure of how far to trust it. If the model is calibrated, so that its stated probabilities match how often it is right, a reviewer can accept its confident answers and check the rest by hand. This paper fine-tunes Qwen3-4B-Base with low-rank adapters and a classification head for evidence inference (Lehman et al., 2019; DeYoung et al., 2020): given a trial's article and a prompt that names an intervention, a comparator and an outcome, classify the result as a significant increase, a significant decrease or no significant difference. The articles of Evidence Inference 2.0 come from the open-access subset of PubMed Central, which includes articles under non-commercial and no-derivatives licenses, so the released model was fine-tuned only on the 1,431 of 2,657 training articles under CC BY, CC0 or the public domain; it is released under the Apache License 2.0 and runs on a laptop. On the test split, its expected calibration error is 0.017 and its log loss is lower than that of Gemma 4 31B-it given the same input zero-shot, meeting criteria registered on the Open Science Framework before testing. Its macro-F1, 0.925, is above every published result on that split that I found, though those systems used other inputs, scoring rules and prompt filters, and the test labels are public. Accepting its 80 percent most confident answers leaves an error rate of 2.6 percent among them. Context ablations show that the answers depend on the article, but the title alone carries some signal, which an informative title or recall of the trial could explain.

  • 1 authors
·
Oct 4

CaliBench: Are the Stochastic Dynamics of Video World Models Physically Calibrated?

Video world models approximate the stochastic distribution of physical outcomes through generative sampling, but existing benchmarks score individual generations or compare distributions coarsely over a whole dataset, leaving the fine-grained aleatoric uncertainty of specific phenomena untested. We introduce CaliBench, which scores outcomes in a physically interpretable discrete space - a bin index, a die face, a suit, a colour - rather than a learned feature space such as in FID, so the distance from a known reference distribution is measured directly. We curate outcome spaces whose reference is known in closed form (binomial Galton boards, Bernoulli forks, uniform dice/cards/lottery, a skewed European-roulette colour), enabling an exact calibration test. We decompose performance into two orthogonal axes that a single accuracy metric conflates: scorability, the fraction of generations yielding a scoreable outcome, and calibration, the total variation distance from the reference on that sample. A chi-squared test assesses significance; as calibration is its null hypothesis it can evidence only miscalibration, and at N=32 per cell detects only large deviations. We apply it to nine scenes and six image-to-video models (WAN-2.7, SeeDance-2.0, HappyHorse-1.0, Veo 3.1, Runway Gen-4.5, Cosmos3-Super), 32 generations each. Models consistently concentrate probability mass on a few outcomes rather than reproducing the reference. Most scene-model combinations are significantly miscalibrated, in the extreme collapsing to one outcome, as Veo 3.1 does on dice. On roulette, generations often leave the ball ambiguously placed, giving several models low scorability. Performance varies by scene: no model dominates all nine. We release the protocol and a metric (mean normalised total variation, mnTV) for comparing new models against our results.

  • 6 authors
·
Aug 16

Calibrated Surprise: An Information-Theoretic Account of Creative Quality

In the era of large language models, creative writing quality lacks a computable theoretical anchor. The dominant approaches are rubric scoring -- decomposing holistic aesthetic judgment into sub-scores -- and RLHF preference signals -- replacing quality with group votes. Both bypass the statistical structure of the text itself. This paper provides an information-theoretic foundation to fill this gap. We propose 'calibrated surprise' as the information-theoretic essence of excellent creative writing. This judgment matches reading intuition and covers its opposite. This literary judgment admits a precise mathematical formulation. Under full-dimensional constraints Y, feasible writing choices are forced into an extremely narrow space. The rare survivors are, from the unconstrained perspective, exactly the least predictable choices. Both are measured precisely by Shannon mutual information I(X;Y) = H(X) - H(X|Y) -- 'calibrated' corresponds to H(X|Y) approaching 0; 'surprising' corresponds to H(X) going high. The subtraction structure of the formula naturally separates 'well-grounded surprise' from 'pure noise'. We use token-level logprobs from Qwen1.5-7B as an operational proxy for the ideal reader's probability distribution. Across 20 pairs (12 Chinese / 8 English) of high-quality vs. systematically degraded literary passages, 20/20 pairs support the core prediction: high-quality passages have systematically higher I(X;Y) than their degraded versions.

BreadStudio Bread Studio
·
Jun 3

What Single-Prompt Accuracy Misses: A Multi-Variant Reliability Audit of Language Models

Single-prompt accuracy is the dominant way to benchmark language models, but it can miss reliability failures that matter. We evaluate a 15-model open-weight corpus, with the main reliability analyses focused on 10 instruct models across five classification and reasoning benchmarks under five prompt variants each, measuring accuracy, token-probability calibration, verbal-confidence calibration, verbal parse rate, and prompt-perturbation spread for every (model x dataset x variant) cell. We find three broad results. First, evaluation design can materially change the conclusion. Switching Expected Calibration Error (ECE) token from a raw to a label-set-normalised definition changes per-cell calibration by a mean absolute 0.149. More strikingly, pairing a chain-of-thought prompt with a first-character evaluator on ARC-Challenge reduces apparent accuracy by 72-88% across all five primary models; two independent repair procedures recover 93.8% and 102.7% of the lost performance, indicating an evaluator-side rather than model-side failure. Second, confidence signals are fragile. On MMLU-Pro, every primary model verbally reports confidence substantially above both its accuracy and its token-probability confidence on the same rows, and verbal parse rate can collapse for a single model on a single prompt variant. Third, prompt robustness does not track parameter count reliably. Across 10 instruct models, the correlation between model size and prompt-perturbation spread ranges from -0.244 to 0.474 across benchmarks. Taken together, these results show that reliability conclusions for small language models depend not only on the model being evaluated, but also on the evaluation pipeline used to measure it. We argue that calibration definitions, evaluator logic, verbal parseability, and prompt robustness should be reported explicitly when making reliability claims.

  • 2 authors
·
May 2

Modeling of learning curves with applications to pos tagging

An algorithm to estimate the evolution of learning curves on the whole of a training data base, based on the results obtained from a portion and using a functional strategy, is introduced. We approximate iteratively the sought value at the desired time, independently of the learning technique used and once a point in the process, called prediction level, has been passed. The proposal proves to be formally correct with respect to our working hypotheses and includes a reliable proximity condition. This allows the user to fix a convergence threshold with respect to the accuracy finally achievable, which extends the concept of stopping criterion and seems to be effective even in the presence of distorting observations. Our aim is to evaluate the training effort, supporting decision making in order to reduce the need for both human and computational resources during the learning process. The proposal is of interest in at least three operational procedures. The first is the anticipation of accuracy gain, with the purpose of measuring how much work is needed to achieve a certain degree of performance. The second relates the comparison of efficiency between systems at training time, with the objective of completing this task only for the one that best suits our requirements. The prediction of accuracy is also a valuable item of information for customizing systems, since we can estimate in advance the impact of settings on both the performance and the development costs. Using the generation of part-of-speech taggers as an example application, the experimental results are consistent with our expectations.

  • 3 authors
·
Feb 4, 2024

Prior Audit-Repair Context Shifts LLM Verifier Thresholds Toward Leniency

Automated checking pipelines increasingly place one language model as the checker and another (or the same one) as the fixer. We ask whether that wiring changes what the checker reports. Measuring false alarms on human-verified-correct ProcessBench traces with the present task held byte-identical, we find that a completed audit -> repair episode already in the model's context lowers false alarms in 15 of 15 model x wording combinations, by 2.8 to 11.5 percentage points against a length-matched non-audit control, a 9 to 25% reduction relative to that control. The direction contradicts what the accumulated-message literature predicts: an episode whose audit reported an error lowers false alarms further still, at all five wordings on the model where that manipulation lands cleanly, though a negativity asymmetry predicts more flagging. Decomposing the episode finds repair content and audit verdict complementary: different components carry the effect on different model families. Signal-detection analysis locates the change in the threshold rather than in discrimination -- the criterion moves in 15 of 15 combinations and survives correction in 13 while d' survives in none, though the d' test is half as sensitive by construction -- and a hand audit of 50 false alarms finds 82% simply wrong, so at this operating point the shift need not be harmful. With reasoning enabled the effect keeps its relative size on both models tested, and the threshold reading holds there too.

  • 2 authors
·
Aug 16 2

Preliminary assessment of a cost-effective headphone calibration procedure for soundscape evaluations

The introduction of ISO 12913-2:2018 has provided a framework for standardized data collection and reporting procedures for soundscape practitioners. A strong emphasis was placed on the use of calibrated head and torso simulators (HATS) for binaural audio capture to obtain an accurate subjective impression and acoustic measure of the soundscape under evaluation. To auralise the binaural recordings as recorded or at set levels, the audio stimuli and the headphone setup are usually calibrated with a HATS. However, calibrated HATS are too financially prohibitive for most research teams, inevitably diminishing the availability of the soundscape standard. With the increasing availability of soundscape binaural recording datasets, and the importance of cross-cultural validation of the soundscape ISO standards, e.g.\ via the Soundscape Attributes Translation Project (SATP), it is imperative to assess the suitability of cost-effective headphone calibration methods to maximise availability without severely compromising on accuracy. Hence, this study objectively examines an open-circuit voltage (OCV) calibration method in comparison to a calibrated HATS on various soundcard and headphone combinations. Preliminary experiments found that calibration with the OCV method differed significantly from the reference binaural recordings in sound pressure levels, whereas negligible differences in levels were observed with the HATS calibration.

  • 7 authors
·
May 10, 2022