2/16/2026
Low-Rank Online Dynamic Assortment with Dual Contextual Information
ML research on dynamic assortment optimization with dual contextual information for e-commerce recommendation systems.
2/16/2026
Continuous-time q-Learning for Jump-Diffusion Models under Tsallis Entropy
ML research on continuous-time Q-learning for jump-diffusion models using Tsallis entropy regularization instead of Shannon entropy.
2/16/2026
SaVe-TAG: LLM-based Interpolation for Long-Tailed Text-Attributed Graphs
SaVe-TAG uses LLM-based interpolation to address long-tailed class imbalance in text-attributed graphs for GNN generalization.
2/16/2026
Measure-to-measure interpolation using Transformers
Theoretical analysis of Transformers as measure-to-measure maps implemented as interacting particle systems on the unit sphere.
2/16/2026
Post-hoc Probabilistic Vision-Language Models
Post-hoc method adding probabilistic uncertainty to vision-language models like CLIP to better handle domain shifts in downstream tasks.
2/16/2026
Hierarchical Retrieval at Scale: Bridging Transparency and Efficiency
Hierarchical retrieval method offering interpretability and efficiency improvements over embedding-based similarity search for large-scale systems.
2/16/2026
PoliCon: Evaluating LLMs on Achieving Diverse Political Consensus Objectives
PoliCon benchmark evaluates LLMs on achieving political consensus objectives using deliberation records from European Parliament.
2/16/2026
How to Train Your LLM Web Agent: A Statistical Diagnosis
Statistical diagnosis and training methods for LLM-based web agents addressing multi-step interactions and reducing post-training compute costs.
2/16/2026
Load Balancing for AI Training Workloads
Technical analysis of load-balancing designs for AI training workloads, comparing approaches and establishing optimality bounds for distributed training.
2/16/2026
Highlight & Summarize: RAG without the jailbreaks
Highlight & Summarize method for RAG systems to prevent jailbreaking and model hijacking of LLMs through prompt injection defense.
2/16/2026
ToolACE-MT: Non-Autoregressive Generation for Agentic Multi-Turn Interaction
ToolACE-MT enables non-autoregressive generation for multi-turn LLM agent interactions with complex function calls, reducing data generation costs.
2/16/2026
Bayesian Ego-graph Inference for Networked Multi-Agent Reinforcement Learning
Bayesian method for decentralized multi-agent reinforcement learning over networked graphs with dynamic neighborhoods and constrained communication.
2/16/2026
HEART: Emotionally-Driven Test-Time Scaling of Language Models
HEART framework uses emotional cues during test-time scaling to improve LLM problem-solving by preventing repetitive thought patterns through alternating critical and encouraging tones.
2/16/2026
The Conditions of Physical Embodiment Enable Generalization and Care
Research on physical embodiment constraints for AI agents in eldercare and disaster response scenarios, addressing generalization and care provision under uncertainty.
2/16/2026
VoiceAgentBench: Are Voice Assistants ready for agentic tasks?
VoiceAgentBench evaluates speech language models on agentic multi-turn tasks and adversarial robustness beyond isolated capabilities.
2/16/2026
TA-KAND: Two-stage Attention Triple Enhancement and U-KAN based Diffusion For Few-shot Knowledge Graph Completion
Knowledge graph completion using attention mechanisms and diffusion models for few-shot learning on long-tailed relations.
2/16/2026
Tuberculosis Screening from Cough Audio: Baseline Models, Clinical Variables, and Uncertainty Quantification
Standardized framework for tuberculosis detection from cough audio combining machine learning models with clinical variables and uncertainty quantification.
2/16/2026
PIDSMaker: Building and Evaluating Provenance-based Intrusion Detection Systems
Framework and tools for building and standardizing provenance-based intrusion detection systems with consistent evaluation metrics.
2/16/2026
WideSeek-R1: Exploring Width Scaling for Broad Information Seeking via Multi-Agent Reinforcement Learning
Width scaling approach for multi-agent systems addressing broad information seeking through organizational capability rather than single-agent depth.
2/16/2026
Reinforced Attention Learning
Policy-gradient framework optimizing internal attention distributions in multimodal LLMs for improved reasoning without verbose rationales.
2/16/2026
Visual concept ranking uncovers medical shortcuts used by large multimodal models
Method for identifying visual concepts in large multimodal models to audit medical decision-making and uncover shortcut behaviors in skin lesion classification.
2/16/2026
FiMI: A Domain-Specific Language Model for Indian Finance Ecosystem
Domain-specialized financial language model for Indian digital payment systems, developed by NPCI using multi-stage training on 68B tokens.
2/16/2026
Learned Finite Element-based Regularization of the Inverse Problem in Electrocardiographic Imaging
Space-time regularization using learned finite element methods for inverse electrocardiographic imaging of cardiac electrical activity.
2/16/2026
Predicting Open Source Software Sustainability with Deep Temporal Neural Hierarchical Architectures and Explainable AI
Deep temporal neural hierarchical model for predicting open source software sustainability using contribution patterns and community metrics.
2/16/2026
GPU-Fuzz: Finding Memory Errors in Deep Learning Frameworks
GPU-Fuzz fuzzer for finding memory errors in deep learning frameworks (PyTorch, TensorFlow, PaddlePaddle) by modeling operator constraints.
2/16/2026
Why Agentic Theorem Prover Works: A Statistical Provability Theory of Mathematical Reasoning Models
Statistical provability theory explaining why agentic theorem provers combining reasoning models with library retrieval, planning, and proof verification achieve strong mathematical reasoning performance.
2/16/2026
Characterizing Trainability of Instantaneous Quantum Polynomial Circuit Born Machines
Theoretical characterization of trainability for instantaneous quantum polynomial circuit born machines as quantum generative models.
2/16/2026
Adaptive Power Iteration Method for Differentially Private PCA
Privacy-preserving algorithm for computing top singular vectors using adaptive power iteration with differential privacy guarantees.
2/16/2026
Free Lunch for Stabilizing Rectified Flow Inversion
Training-free inversion stabilization for rectified-flow generative models, improving reconstruction and editing tasks without additional training.
2/16/2026
When Should LLMs Be Less Specific? Selective Abstraction for Reliable Long-Form Text Generation
Selective Abstraction framework for reducing factual errors in LLM long-form generation by enabling partial uncertainty-based abstention instead of binary all-or-nothing approach.
2/16/2026
The Implicit Bias of Logit Regularization
Theoretical analysis of logit regularization in linear classifiers, examining implicit bias mechanisms of label smoothing and related convex penalties.
2/16/2026
T3D: Few-Step Diffusion Language Models via Trajectory Self-Distillation with Direct Discriminative Optimization
Trajectory self-distillation method for improving few-step decoding in diffusion language models, enabling faster parallel token generation with maintained quality.
2/16/2026
A Lightweight LLM Framework for Disaster Humanitarian Information Classification
arXiv paper on lightweight, parameter-efficient fine-tuning framework for disaster tweet classification using LLMs in resource-constrained settings.
2/16/2026
From Biased Chatbots to Biased Agents: Examining Role Assignment Effects on LLM Agent Robustness
arXiv case study examining how demographic-based persona assignments affect LLM agent robustness and task performance, highlighting operational risks beyond text generation bias.
2/16/2026
Retrieval-Augmented Self-Taught Reasoning Model with Adaptive Chain-of-Thought for ASR Named Entity Correction
arXiv paper proposing retrieval-augmented self-taught reasoning model with adaptive chain-of-thought for correcting named entity errors in automatic speech recognition using LLMs.
2/16/2026
Grandes Modelos de Linguagem Multimodais (MLLMs): Da Teoria \`a Pr\'atica
arXiv paper on multimodal LLMs covering fundamentals, emblematic models, and practical techniques for preprocessing, prompt engineering, and building pipelines with LangChain and LangGraph.
2/16/2026
propella-1: Multi-Property Document Annotation for LLM Data Curation at Scale
propella family of small multilingual LLMs annotating documents across 18 properties for flexible LLM pretraining data curation.
2/16/2026
RankLLM: Weighted Ranking of LLMs by Quantifying Question Difficulty
RankLLM framework quantifying question difficulty in benchmarks to better differentiate LLM capabilities across model comparisons.
2/16/2026
RBCorr: Response Bias Correction in Language Models
RBCorr method to correct response bias in language models' performance on fixed-response evaluation questions.
2/16/2026
Unleashing Low-Bit Inference on Ascend NPUs: A Comprehensive Evaluation of HiFloat Formats
Evaluation of HiFloat low-bit formats for efficient LLM inference on Ascend NPUs with precision-efficiency comparisons.
2/16/2026
CLASE: A Hybrid Method for Chinese Legalese Stylistic Evaluation
CLASE hybrid method for evaluating stylistic quality of legal text generated by LLMs without manual expert annotation.
2/16/2026
Beyond Normalization: Rethinking the Partition Function as a Difficulty Scheduler for RLVR
Reinterprets partition function in GFlowNet-based RL for LLMs as difficulty scheduler to improve reasoning diversity while maintaining performance.
2/16/2026
Learning Ordinal Probabilistic Reward from Preferences
Novel reward modeling paradigm combining generative and discriminative approaches for probabilistic preference-based LLM alignment.
2/16/2026
$\mathcal{X}$-KD: General Experiential Knowledge Distillation for Large Language Models
Experiential Knowledge Distillation framework for compressing LLMs by learning from teacher model's original training environment.
2/16/2026
MedXIAOHE: A Comprehensive Recipe for Building Medical MLLMs
MedXIAOHE medical multimodal foundation model with entity-aware pretraining for clinical image and text understanding tasks.
2/16/2026
ReFilter: Improving Robustness of Retrieval-Augmented Generation via Gated Filter
ReFilter method improving robustness of retrieval-augmented generation systems through gated filtering mechanisms for LLM integration.
2/16/2026
Lamer-SSL: Layer-aware Mixture of LoRA Experts for Continual Multilingual Expansion of Self-supervised Models without Forgetting
Lamer-SSL framework using Layer-Aware Mixture of LoRA Experts for multilingual expansion of speech models without catastrophic forgetting.
2/16/2026
Towards a Diagnostic and Predictive Evaluation Methodology for Sequence Labeling Tasks
Error analysis methodology for evaluating and improving sequence labeling systems with diagnostic and predictive insights.
2/16/2026
Aspect-Based Sentiment Analysis for Future Tourism Experiences: A BERT-MoE Framework for Persian User Reviews
BERT-MoE framework for aspect-based sentiment analysis on Persian tourism reviews addressing low-resource language challenges.
2/16/2026
RAT-Bench: A Comprehensive Benchmark for Text Anonymization
RAT-Bench benchmark for evaluating text anonymization tools' effectiveness at removing PII from data used with LLMs.