On-device research index

arXiv research

A locally-built, LLM-digested index of recent arXiv papers in quant finance, geometry/topology, and statistical ML — keyword search served straight from SQLite on this machine.

168,742 papers · 148 categories

Trend · papers per month

count % of corpus

12 results for “logit-based”

AdaDetectGPT: Adaptive Detection of LLM-Generated Text with Statistical Guaranteescs.CL

AdaDetectGPT improves text authorship detection with statistical guarantees.

problem Determining if text is authored by a human or an LLM.

method Adaptive learning of a witness function from training data to enhance logits-based detectors.

result AdaDetectGPT nearly uniformly improves text authorship detection, with up to 37% improvement.

text authorship LLM detection statistical guarantees logits-based detectors adaptive learning

2025-09-29abs ↗pdf ↗

Machine Unlearning: Linear Filtration for Logit-based Classifierscs.LG

Linear filtration helps delete training data from models.

problem Deleting training data from models when individuals request it.

method Linear filtration as a computationally efficient sanitization method.

result Demonstrates benefits in an adversarial setting over naive deletion schemes.

machine unlearning logit-based classifiers deep neural networks data sanitization classification models

2020-02-07abs ↗pdf ↗

A Semantic-Sampling Framework for Evaluating Calibration in Open-Ended Question Answeringcs.CL

A new framework evaluates LLM calibration in open-ended QA.

problem Evaluating LLM calibration in open-ended QA settings.

method Sem-ECE framework: sampling answers, grouping by semantic classes, and using frequencies as confidence.

result Sem-ECE estimators are unbiased and Sem

_2

achieves smaller calibration error.

LLM calibration open-ended QA semantic sampling expected calibration error logit-based metrics

2026-05-08abs ↗pdf ↗

PLD: A Choice-Theoretic List-Wise Knowledge Distillationcs.LG

PLD distills knowledge using choice-theoretic Plackett-Luce model.

problem Model compression and knowledge transfer between large and small networks.

method PLD uses a weighted list-wise ranking loss based on the Plackett-Luce model.

result PLD achieves consistent gains across diverse architectures and distillation methods.

knowledge distillation logit-based choice-theoretic Plackett-Luce cognitive ranking

2025-06-14abs ↗pdf ↗

Beyond Heuristic Tuning: Power-Calibrated LLM Watermarkingstat.ML

Develops a power-calibrated framework for LLM watermarking, optimizing tradeoffs between detectability and distortion.

problem The trade-off between detectability and semantic distortion in logit-based watermarking.

method Power-calibrated statistical framework for watermark hyperparameters, establishing explicit relationships.

result Derives practical parameter selection procedures achieving optimal tradeoffs under constraints.

watermarking LLM logit detectability semantic distortion

2026-07-06abs ↗pdf ↗

Learning to Scale Logits for Temperature-Conditional GFlowNetscs.LG

Logit-GFN accelerates GFlowNets training by scaling logits based on temperature.

problem Training temperature-conditional GFlowNets is numerically challenging.

method Logit-GFN uses a learned function of temperature to scale policy logits.

result Logit-GFN greatly accelerates GFlowNets training and improves generalization and mode discovery.

GFlowNets temperature logits scaling exploration

2023-10-04abs ↗pdf ↗

Energy-based Epistemic Uncertainty for Graph Neural Networkscs.LG

GEBM improves uncertainty quantification in graph neural networks.

problem Challenges in quantifying epistemic uncertainty in graph neural networks.

method Energy-based model (EBM) that aggregates uncertainty at different structural levels.

result Significantly improves predictive robustness and achieves best separation of in-distribution and out-of-distribution data.

Graph Neural Networks Uncertainty Quantification Energy-Based Models Epistemic Uncertainty Graph Diffusion

2024-06-06abs ↗pdf ↗

Sampling Control for Imbalanced Calibration in Semi-Supervised Learningcs.LG

Unified framework suppresses model bias in semi-supervised learning with decoupled sampling control.

problem Class imbalance in semi-supervised learning, especially with distributional mismatches.

method Unified framework SC-SSL with decoupled sampling control, explicit expansion capability, and adaptive sampling probabilities.

result Consistent and state-of-the-art performance across various benchmark datasets and distribution settings.

semi-supervised learning class imbalance model bias sampling control distributional mismatches

2025-11-24abs ↗pdf ↗

MANO: Exploiting Matrix Norm for Unsupervised Accuracy Estimation Under Distribution Shiftscs.LG

MANO normalizes logits to estimate test accuracy without labels.

problem Estimating test accuracy of OOD samples without labels.

method Applies

L_p

norm to normalized logits.

result Achieves state-of-the-art performance across various architectures.

logits unsupervised accuracy estimation distribution shift matrix norm

2024-05-29abs ↗pdf ↗

FIFA: Making Fairness More Generalizable in Classifiers Trained on Imbalanced Datacs.LG

FIFA improves fairness in imbalanced datasets by encouraging both classification and fairness generalization.

problem Imbalanced datasets lead to poor fairness generalization in classifiers.

method FIFA: Imbalance-Fairness-Aware approach that encourages both classification and fairness generalization.

result FIFA improves fairness generalization on real-world datasets.

fairness imbalanced data classification generalization FIFA

2022-06-06abs ↗pdf ↗

Estimation, Prediction, and Assortment Optimization for Markov Chain Choice Models with Panel Datacs.LG

Proposes new methods for Markov chain choice models with panel data.

problem Dependence among transactions for the same customer in historical data.

method Expectation-maximization (EM) algorithms incorporating partial-ordering preference information.

result EM algorithms outperform traditional methods on synthetic and real datasets.

Markov chain choice models panel data EM algorithms partial-ordering customer preference

2026-07-10abs ↗pdf ↗

Improved Balanced Classification with Theoretically Grounded Loss Functionscs.LG

This paper introduces new loss functions for balanced multi-class classification.

problem Balancing class imbalance in multi-class classification.

method Introduces two new surrogate loss families: GLA and GCA.

result GCA losses offer stronger theoretical guarantees in imbalanced settings.

balanced classification loss functions multi-class class imbalance theoretical guarantees

2025-12-30abs ↗pdf ↗