Formal analysis of algorithmic learning (grokking) in infinite-width transformers with computational complexity bounds and generalization theory.
Speech recognition system for Huntington's disease detection using clinical speech corpus and ASR biomarker supervision.
Bayesian optimization method combining probabilistic models with hybrid systems for efficient black-box optimization of expensive functions.
Parameter-efficient fine-tuning approach for continual learning that controls representation-level adaptation in pretrained models.
Machine unlearning method using reference-guided approach to remove data influence from trained models while preserving utility.
DISCOMAX: differentiable physics-consistent ML method for predicting thermodynamic phase equilibria.
Method for out-of-distribution detection in single-domain models using teacher-guided training to avoid domain-sensitivity collapse.
Duration-aware scheduling policy for ASR serving pipelines addressing head-of-line blocking under workload drift.
Benchmark for evaluating state space models on long-sequence biological imaging data with sparse, stochastic temporal processes.
Federated learning method conditioning global model on client-specific PCA statistics for heterogeneous data without extra communication.
Heavy-tailed PCA variant for robust dimensionality reduction on data with impulsive noise and heavy-tailed distributions.
Hindsight-Anchored Policy Optimization for sparse-reward RL in reasoning model post-training, addressing advantage collapse.
MR-Search: meta-reinforcement learning framework for agentic search with self-reflection, enabling agents to improve in-context exploration.
Analysis of adversarial attacks on LLMs showing polynomial-exponential crossover in success rates with inference-time samples.
Teleodynamic Learning: proposed paradigm treating intelligence as coupled evolution of representation, adaptation, and resource constraints.
Machine learning and transformer models for financial fraud detection in Bangla and English languages.
Open-source Python simulation package for modeling antibiotic prescribing and antimicrobial resistance dynamics in RL-compatible environment.
Longitudinal active feature acquisition method for optimizing predictive performance when measurements incur cost or risk.
Offline reinforcement learning with digital twin verification for automating mechanical ventilation settings in ICU.
Manifold learning approach using Finsler geometry to capture asymmetric information in high-dimensional data embeddings.
Neural method for stable estimation of statistical dependence in autoencoders using orthonormal density-ratio decomposition.
Zero-shot learning approach for automatically detecting semantic column types in relational tables without labeled training data.
Unsupervised neural combinatorial optimization method (UniHetCO) for multi-problem learning on graph subset-selection problems without ground-truth solutions.
Dual-path approach combining discrete mark prediction and continuous dynamics for marked temporal point processes using neural ODEs.
Proximal relaxation method for improving nonlinear probabilistic latent variable models in soft sensor applications, addressing training inefficiencies.
Deep learning network-temporal models for multivariate time series traffic prediction, addressing topological interdependency and complex temporal patterns.
Sorometry pipeline for automated phytolith analysis using AI to digitize and classify microscope images, replacing manual labor-intensive analysis.
Neuro-symbolic VLM agents for time series event detection using natural language descriptions, addressing semantic event classification with scarce labeled data.
Theoretical analysis proving attention sinks are necessary in softmax transformers for trigger-conditional tasks, formalizing why attention collapses to content-agnostic positions.
KEPo analyzes poisoning attacks on Graph-RAG systems where attackers inject malicious texts to manipulate LLM outputs.
Sharpness-aware minimization for stable item embedding learning in federated recommendation systems preserving privacy.
LongFlow compresses KV cache for reasoning models like o1 and R1, reducing memory and bandwidth during long output generation.
Conditional feature disentanglement approach for user-controllable privacy in wearable sensor-based human activity recognition.
Multi-Task Anti-Causal learning framework exploiting cross-task invariances to infer latent causes from observed urban event reports.
CAETC method using adversarial autoencoding for counterfactual estimation with time-dependent confounding in observational data.
Integrates survival analysis with classification for early chronic disease risk prediction using EMR data.
H-EARS combines potential-based reward shaping with energy-aware regularization for efficient deep reinforcement learning control.
AutoScout automates ML system configuration via structured optimization over model parallelism, communication, and runtime parameters.
Investigates partial RoPE rotations in transformers, reducing memory at long contexts while maintaining performance.
Personalized federated learning using Gaussian generative modeling to handle data heterogeneity across distributed clients.
Studies continual RL for Vision-Language-Action models, finding sequential fine-tuning avoids catastrophic forgetting without complex strategies.
Neuromodulated constrained autoencoders for context-dependent dimensionality reduction in varying environments.
Policy gradient methods for LLM reasoning naturally reduce trajectory diversity; proposes entropy-preserving training approach.
EvoFlows model for protein engineering using edit-based flow-matching to predict mutations on template sequences.
Examines calibration's role in reducing predictive multiplicity and improving stability in high-stakes ML classifier deployments.
Social bandit learning framework combining individual and collective learning in populations of reinforcement learning agents.
Theoretical study of Follow-the-Perturbed-Leader algorithm optimality in semi-bandit problems with best-of-both-worlds guarantees.
Analyzes model collapse when LLM-generated text re-enters training data as data consumption grows, proposing replay-based solutions.
Method for agents to autonomously discover symmetry groups for disentangled representation learning without prior structural knowledge.
Unifies membership inference attacks (LiRA, RMIA, BASE) as instances of exponential-family statistical framework for model privacy auditing.