4/3/2026
ORBIT: Scalable and Verifiable Data Generation for Search Agents on a Tight Budget
ORBIT generates 20K training queries for search agents integrating LMs with web search using scalable and verifiable methods.
4/3/2026
DR-LoRA: Dynamic Rank LoRA for Fine-Tuning Mixture-of-Experts Models
DR-LoRA assigns dynamic ranks to expert modules in MoE models for efficient parameter-specific fine-tuning of LLMs.
4/3/2026
Cognitive Friction: A Decision-Theoretic Framework for Bounded Deliberation in Tool-Using Agents
Triadic Cognitive Architecture for tool-using agents with principled bounds on information-acquisition costs and deliberation.
4/3/2026
BIOGEN: Evidence-Grounded Multi-Agent Reasoning Framework for Transcriptomic Interpretation in Antimicrobial Resistance
BIOGEN multi-agent reasoning framework using evidence-grounding for transcriptomic interpretation in antimicrobial resistance.
4/3/2026
TaCarla: A comprehensive benchmarking dataset for end-to-end autonomous driving
TaCarla comprehensive benchmarking dataset for end-to-end autonomous driving with perception and planning tasks.
4/3/2026
MemFactory: Unified Inference & Training Framework for Agent Memory
MemFactory unified framework for training and inference in memory-augmented LLMs using RL to optimize memory operations.
4/3/2026
Sven: Singular Value Descent as a Computationally Efficient Natural Gradient Method
Sven optimization algorithm exploiting natural loss decomposition using Moore-Penrose pseudoinverse for efficient neural network training.
4/3/2026
Forecasting Supply Chain Disruptions with Foresight Learning
Framework training LLMs to forecast supply chain disruptions using calibrated probabilistic forecasts from disruption outcomes.
4/3/2026
Model Merging via Data-Free Covariance Estimation
Principled layer-wise optimization approach for model merging via data-free covariance estimation without task-specific training.
4/3/2026
Malliavin Calculus for Counterfactual Gradient Estimation in Adaptive Inverse Reinforcement Learning
Novel Langevin-based algorithm for adaptive inverse reinforcement learning using Malliavin calculus for gradient estimation.
4/3/2026
PI-JEPA: Label-Free Surrogate Pretraining for Coupled Multiphysics Simulation via Operator-Split Latent Prediction
PI-JEPA: Physics-informed surrogate model for multiphysics simulation exploiting unlabeled parameter fields via latent prediction.
4/3/2026
Residuals-based Offline Reinforcement Learning
Residuals-based offline reinforcement learning approach for high-stakes applications with restrictive data coverage assumptions.
4/3/2026
Test-Time Scaling Makes Overtraining Compute-Optimal
Train-to-Test scaling laws optimizing model size, training tokens, and inference samples jointly for compute-optimal LLM deployment.
4/3/2026
When Reward Hacking Rebounds: Understanding and Mitigating It with Representation-Level Signals
Study of reward hacking in LLM RL showing reproducible failure patterns and mitigation strategies using representation-level signals.
4/3/2026
DISCO-TAB: A Hierarchical Reinforcement Learning Framework for Privacy-Preserving Synthesis of Complex Clinical Data
Hierarchical RL framework for privacy-preserving synthetic clinical data generation combining LLMs with structured learning.
4/3/2026
CuTeGen: An LLM-Based Agentic Framework for Generation and Optimization of High-Performance GPU Kernels using CuTe
CuTeGen: LLM-based agentic framework for automated generation and optimization of high-performance GPU kernels using CuTe abstraction.
4/3/2026
Matching Accuracy, Different Geometry: Evolution Strategies vs GRPO in LLM Post-Training
Comparative study of Evolution Strategies vs GRPO for LLM post-training showing ES achieves comparable accuracy with different parameter geometry.
4/3/2026
Care-Conditioned Neuromodulation for Autonomy-Preserving Supportive Dialogue Agents
Care-Conditioned Neuromodulation framework for LLM-based dialogue agents that balances helpfulness with user autonomy preservation.
4/3/2026
Learning from the Right Rollouts: Data Attribution for PPO-based LLM Post-Training
Influence-Guided PPO framework for LLM post-training that filters noisy rollouts using data attribution to improve training efficiency.
4/3/2026
Training In-Context and In-Weights Mixtures Via Contrastive Context Sampling
Research on training LLMs to develop both in-context and in-weights learning capabilities simultaneously via contrastive context sampling.
4/3/2026
Pseudo-Quantized Actor-Critic Algorithm for Robustness to Noisy Temporal Difference Error
Novel reinforcement learning algorithm addressing noisy temporal difference errors in deep RL through pseudo-quantization methods.
4/3/2026
Expert-Choice Routing Enables Adaptive Computation in Diffusion Language Models
arXiv paper on expert-choice routing for diffusion language models. Deterministic load balancing improves throughput and convergence vs token-choice.
4/3/2026
CRIT: Graph-Based Automatic Data Synthesis to Enhance Cross-Modal Multi-Hop Reasoning
arXiv paper on CRIT, graph-based automatic data synthesis for cross-modal multi-hop reasoning. Generates complementary image-text data.
4/3/2026
Label Shift Estimation With Incremental Prior Update
arXiv paper on label shift estimation with incremental prior updates. Addresses distribution mismatch between training and deployment.
4/3/2026
Coupled Query-Key Dynamics for Attention
arXiv paper on coupled query-key dynamics for scaled dot-product attention. Improves language modeling perplexity by 6-7% on WikiText-103.
4/3/2026
MiCA Learns More Knowledge Than LoRA and Full Fine-Tuning
arXiv paper introducing MiCA, parameter-efficient LLM fine-tuning method adapting minor singular vector subspaces. Outperforms LoRA on knowledge retention.
4/3/2026
Transformer self-attention encoder-decoder with multimodal deep learning for response time series forecasting and digital twin support in wind structural health monitoring
arXiv paper on transformer encoder-decoder with multimodal learning for wind structural health monitoring and digital twins.
4/3/2026
MATA-Former & SIICU: Semantic Aware Temporal Alignment for High-Fidelity ICU Risk Prediction
arXiv paper on MATA-Former for ICU risk prediction using semantic-aware temporal alignment. Clinical-logic-aligned transformer architecture.
4/3/2026
Koopman-Based Nonlinear Identification and Adaptive Control of a Turbofan Engine
arXiv paper applying Koopman operator methods for multivariable control of turbofan engines. Meta-heuristic extended dynamic mode decomposition.
4/3/2026
DDCL: Deep Dual Competitive Learning: A Differentiable End-to-End Framework for Unsupervised Prototype-Based Representation Learning
arXiv paper on DDCL, differentiable end-to-end framework for unsupervised prototype-based representation learning. Integrates feature learning with clustering.
4/3/2026
FourierMoE: Fourier Mixture-of-Experts Adaptation of Large Language Models
arXiv paper proposing FourierMoE for parameter-efficient LLM fine-tuning in multi-task settings. Mixture-of-experts approach addressing task interference.
4/3/2026
Bridging Deep Learning and Integer Linear Programming: A Predictive-to-Prescriptive Framework for Supply Chain Analytics
arXiv paper integrating deep learning forecasting with integer linear programming for supply chain analytics. Three-stage framework on 180K transactions.
4/3/2026
Graph Neural Operator Towards Edge Deployability and Portability for Sparse-to-Dense, Real-Time Virtual Sensing on Irregular Grids
arXiv paper on graph neural operators for real-time virtual sensing on irregular grids with edge deployment. Addresses sparse-to-dense field inference.
4/3/2026
Physics Informed Reinforcement Learning with Gibbs Priors for Topology Control in Power Grids
arXiv paper combining reinforcement learning with physics-informed priors (Gibbs) for power grid topology control. Semi-Markov framework.
4/3/2026
CANDI: Curated Test-Time Adaptation for Multivariate Time-Series Anomaly Detection Under Distribution Shift
arXiv paper on test-time adaptation for multivariate time-series anomaly detection under distribution shift. Curated adaptation framework (CANDI).
4/3/2026
Towards Intrinsically Calibrated Uncertainty Quantification in Industrial Data-Driven Models via Diffusion Sampler
arXiv paper introducing diffusion-based posterior sampling for uncertainty quantification in industrial data-driven models. Addresses safety-critical applications.
4/3/2026
Robust Graph Representation Learning via Adaptive Spectral Contrast
arXiv paper on spectral graph contrastive learning addressing high-frequency signal variance in heterophilic graphs. Theoretical analysis with regret bounds.
4/3/2026
DDCL-INCRT: A Self-Organising Transformer with Hierarchical Prototype Structure (Theoretical Foundations)
Self-organising transformer with hierarchical prototype structure that automatically determines optimal architecture dimensions (heads, depth, width) during training.
4/3/2026
Enhancing the Reliability of Medical AI through Expert-guided Uncertainty Modeling
Expert-guided uncertainty modeling for medical AI systems to improve reliability and enable human experts to prioritize high-risk diagnostic cases.
4/3/2026
The Rank and Gradient Lost in Non-stationarity: Sample Weight Decay for Mitigating Plasticity Loss in Reinforcement Learning
Theoretical analysis of plasticity loss in deep reinforcement learning due to non-stationarity, proposing sample weight decay mitigation technique.
4/3/2026
PAC-Bayesian Reward-Certified Outcome Weighted Learning
PAC-Bayesian framework for outcome weighted learning that incorporates reward uncertainty into policy selection with finite-sample guarantees.
4/3/2026
annbatch unlocks terabyte-scale training of biological data in anndata
annbatch: Mini-batch loader for terabyte-scale biological data training in anndata format, addressing memory bottlenecks in ML pipelines for bioinformatics.
4/3/2026
Learn by Surprise, Commit by Proof
LSCP: Self-gated post-training framework for autonomous knowledge acquisition using self-generated Q&A chains and adaptive learning rates based on model conviction.
4/3/2026
Generalization Bounds and Statistical Guarantees for Multi-Task and Multiple Operator Learning with MNO Networks
Statistical analysis of multi-task and multiple operator learning architectures with generalization bounds and theoretical guarantees.
4/3/2026
World Action Verifier: Self-Improving World Models via Forward-Inverse Asymmetry
Self-improving world models using forward-inverse asymmetry to improve robustness across suboptimal actions for policy evaluation and planning.
4/3/2026
Apriel-Reasoner: RL Post-Training for General-Purpose and Efficient Reasoning
RL post-training framework for building general-purpose reasoning models across diverse domains with verifiable rewards, addressing multi-domain optimization challenges.
4/3/2026
Feature Weighting Improves Pool-Based Sequential Active Learning for Regression
Active learning method using feature weighting for regression tasks to optimize sample selection from unlabeled pools under budget constraints.
4/3/2026
Ouroboros: Dynamic Weight Generation for Recursive Transformers via Input-Conditioned LoRA Modulation
ArXiv paper on dynamic weight generation for recursive transformers using input-conditioned LoRA modulation controller.
4/3/2026
AA-SVD : Anchored and Adaptive SVD for Large Language Model Compression
ArXiv paper on fast SVD-based compression for large language models without retraining, addressing distribution shifts.
4/3/2026
LEO: Graph Attention Network based Hybrid Multi Sensor Extended Object Fusion and Tracking for Autonomous Driving Applications
ArXiv paper on graph attention network for multi-sensor object fusion and tracking in autonomous driving.