Every paper checked — nothing hidden

All papers

Every paper checked in the week of 12–18 Oct 2026 · Small and efficient models — nothing is hidden. Brief papers first.

All2,430 In brief3 More (on-topic)455 Not relevant1,957 Not assessed6 Older27 Could not confirm6 Blogs with no paper16 All blogs19

Paper	Topic	Relevance
Hierarchical Multi-Agent Reasoning Systems for Knowledge-Grounded Material Substitution The paper introduces a hierarchical multi-agent reasoning system combining retrieval and knowledge graphs to find chemical material substitutes. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Focuses on scientific agent workflows, not model compression or efficient inference.
On physical implementability of reinforcement-learning-driven reasoning The paper presents a theoretical negative result on physically implementing reinforcement-learning-driven reasoning in certain physical neural networks. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Theoretical physics paper on hardware stability, not efficient model compression or serving.
A Stochastic Analysis Perspective on Diffusion Models The paper provides a stochastic analysis perspective linking score matching, Langevin sampling, and diffusion model time reversal. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Covers stochastic analysis of diffusion processes, not model distillation or compression.
Embedding Graph Kernel Entropy (EGKE): Fine-grained Hallucination Detection for LLMs from Embeddings The paper proposes an embedding-based graph kernel entropy method for computationally efficient hallucination detection in language models. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Uses sentence embeddings for entropy calculation, not small model design or compression.
Learning Non-Gaussian SEMs with Conditional Risk, and Why the Direction of Recursion Matters The paper extends recursive linear structural equation model identification to non-Gaussian noise using conditional prediction risk. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Presents statistical causality algorithms, not small or efficient neural models.
Do Artificial Neural Networks Need Dendritic Nanotubes? Two Reasons for Shrinking Importance Estimates in Continual Learning The paper examines importance estimate shrinkage to mitigate catastrophic forgetting in regularization-based continual learning. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Addresses continual learning parameter importance, not model compression or efficient inference.
Variance Fingerprints: Component-Wise Identification of Content Factors by Augmentation Design The paper establishes theoretical conditions for identifying individual content factors in self-supervised learning via designed augmentation strengths. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Addresses factor identification in self-supervised models, not model efficiency or compression.
Layer-wise Input Recoverability in Deep Neural Networks The paper probes convolutional neural networks to measure how much layer input information can be reconstructed from layer outputs. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Investigates input recoverability in CNNs, not model compression or efficient serving.
Diagnosing Generalization Boundaries in AI-Based Vulnerability Detection via Calibrated Embedding Distances The paper evaluates embedding distance metrics to detect generalization boundaries and distribution shifts in AI-based code vulnerability models. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Focuses on model representation geometry and degradation, not model compression or efficiency.
Learning to See through Action: Predictive Curiosity for Robots The paper proposes an active robot visual learning framework driven by self-supervised curiosity and world model prediction errors. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Addresses curiosity-driven robotic learning, not efficient model inference or compression.
Decentralized Bandit Learning of Markov Perfect Equilibria The paper proves convergence guarantees for decentralized bandit learning in infinite-horizon, general-sum Markov games. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Addresses continuous Markov game theory, not model efficiency or distillation.
STREGO: A Scalable Implementation of a Trust-Region Framework for Noisy Bayesian Optimization The paper presents a trust-region Bayesian optimization framework designed for noisy, high-dimensional blackbox problems. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Focuses on Bayesian optimization algorithms, not efficient model serving or compression.
Continuous-Time Cloth Dynamics with Physics-Informed Neural ODEs The paper presents a continuous-time neural ODE solver for simulating cloth dynamics at arbitrary timestamps. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Concerns physical cloth simulation solvers, not small language models or model compression.
Class-Conditional Conformal Prediction for Signal Temporal Logic Inference with Many Classes The paper presents a conformal prediction method to reliably infer Signal Temporal Logic formulas for multiclass time series classification. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Does not address model compression, quantization, or efficient inference.
The Sample Complexity of Multicalibration with an ERM Oracle The paper bounds the sample and query complexity of multicalibration using an empirical risk minimization oracle. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Addresses learning theory bounds, not small or efficient model compression.
ConformalGuard: False-Positive-Controlled Action Gating for LM Classifiers The paper proposes a conformal calibration layer to provide finite-sample control over false positives in language model classification workflows. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Proposes a conformal calibration layer, not model compression or serving speedup.
History, Adaptation, and Hidden Gain: Two Prospective Tests of Learned Predictive Control in a Synthetic Neural Population The paper tests predictive neural control against classical controllers in simulated spiking neural populations under varying sensitivity conditions. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Focuses on synthetic neural population control, not small efficient language models.
DPAL: A Test-Time Adaptation Loss for Financial Time Series Forecasting The paper presents an adaptation loss incorporating return direction and price trajectory calibration for test-time financial time series forecasting. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Addresses test-time adaptation loss, not model distillation or serving efficiency.
Decomposing A Neural Learner The paper proposes a probabilistic decomposition of neural models that separates predictable features from an input-dependent reliability score. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Addresses neural interpretability and reliability scoring, not model compression.
Clinical value claims in multimodal AI outpace supporting evidence The paper audits multimodal clinical machine learning studies, finding that incomplete unimodal baselines frequently inflate claimed gains. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Critiques clinical validation literature, not efficient model compression or serving.
Sparse-TD-JEPA: Improving Zero-Shot Reinforcement Learning With Sparse Latent Task Representations The paper applies a sparsity penalty to task representations to sharpen reward signals in zero-shot reinforcement learning. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Investigates zero-shot RL latent representations, not efficient model compression.
A Gated Catchment Hydrology Model (GCHM) for Physically-Interpretable Rainfall-Evapotranspiration-Runoff Simulation Across Diverse Hydroclimatic Regimes The paper presents a mass-conserving neural network architecture designed for interpretable rainfall, evapotranspiration, and runoff simulation across diverse river catchments. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Focuses on hydrological modeling, not small efficient language models.
DEFT: Disentanglement-Enhanced Fine-Tuning for EEG Foundation Models The paper introduces an adapter method that disentangles subject identity from task content during downstream fine-tuning of brainwave foundation models. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Focuses on EEG representation disentanglement, not small or efficient models.
Identication and Estimation of Nonsymmetric Determinantal Point Processes The paper develops identification theory and asymptotic estimation methods for nonsymmetric determinantal point processes capable of modeling positive dependence. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Focuses on nonsymmetric point process estimation, not model efficiency.
Learn the Dynamics, Enforce the Physics: Generative Latent Operators for High-Dimensional Beam Evolution The paper presents a physics-constrained generative latent operator framework to simulate multiscale particle accelerator beam dynamics efficiently. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Surrogate model for particle physics, not small language models or efficient LLM inference.
Chunk-wise Model-based Policy Optimization The paper introduces a chunk-wise model-based reinforcement learning approach that uses single-pass rollouts to optimize policies without recurrent backpropagation through time. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Focuses on RL rollout efficiency rather than small model compression.
Strong Gaussian Approximation of Constant-Learning-Rate SGD: Bias and Trajectory Inference The paper derives strong Gaussian approximations for constant learning rate stochastic gradient descent under heavy-tailed noise to enable trajectory-level inference. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Focuses on theoretical SGD trajectory bounds, not model efficiency techniques.
LLM-Augmented Bayesian Optimization with Adaptive Policies for Chiplet Design Space Exploration The paper combines language models with Bayesian optimization to guide architectural design exploration for heterogeneous chiplet systems. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Optimizes chiplet hardware layout rather than model efficiency or serving.
From Rainfall to Road-Network Functionality: Transferable Graph Forecasting for Flood-Disrupted Transportation Networks The paper proposes a transferable spatiotemporal graph learning framework to forecast road network disruption caused by extreme rainfall across unseen regions. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Addresses spatiotemporal graph learning, not efficient model compression or inference.
TagBO: LLM-Driven Task-Aware Graph Bayesian Optimization for Scientific Discovery The paper introduces a Bayesian optimization method that combines language model relational priors with structural descriptors for mixed-variable scientific discovery. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Focuses on scientific search spaces, not efficient serving or quantization.
Ascent Conditions for Length-Normalized Single-Rollout Policy Updates The paper analyzes policy gradient updates that divide trajectory scores by length, identifying mathematical conditions under which expected returns can decrease. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Covers policy update mathematical theory, not model efficiency or compression.
Triage-PiEvo: Evidence Admission for Principle-Evolving Scientific Agents The paper presents an evidence admission framework that verifies anomalies before updating the persistent principles of scientific language model agents. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Deals with autonomous scientific agents, not model compression or efficiency.
One-Way Coupling via Patchwise Adaptive Normalization in Transformer Surrogates The paper proposes patchwise adaptive normalization to represent one-way coupled exogenous fields in transformer surrogates for partial differential equations. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Deals with scientific PDE surrogates, not general language model compression.
Understanding Hidden Spatial Preferences in Bayesian Optimization This paper identifies spatial boundary biases in Bayesian optimization and modifies exploration using parameterized spatial weight fields. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Studies spatial preferences in Bayesian optimization, not model compression.
Multi-Token Random-Walk Learning via Meeting-Triggered Information Exchange This paper analyzes the complexity and theoretical benefits of meeting-triggered information exchange in multi-token random-walk decentralized learning. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Studies distributed training algorithms, not model compression or efficient serving.
VitaBench: Memory Lifecycles Beyond Retrieval This paper introduces a benchmark to evaluate agent memory lifecycle decisions, including compression, consolidation, and deletion. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Presents a memory evaluation benchmark, not model compression or efficient serving.
More Domains, No Gains? FLO: Fast-Slow Reciprocal Learning for Multi-Domain Graph Pre-training This paper studies why multi-domain graph pre-training fails to generalize and proposes a fast-slow reciprocal learning framework. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Studies multi-domain graph pre-training, not model compression or serving.
Random Fourier Features in Near-Linear Time This paper provides a tighter analysis enabling near-linear time computation of random Fourier feature maps with near-subgaussian concentration. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Focuses on kernel feature computation, not small models or inference.
Rethinking Opinion Measurement Interfaces for Human-in-the-Loop Optimization This paper evaluates human opinion measurement interfaces and recommends listwise preference collection for human-in-the-loop optimization. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Analyzes human-in-the-loop interfaces, not model compression or efficient serving.
Formalizing Forgetting: Theory of Forgetting for Online Continual Learning This paper provides theoretical upper bounds on catastrophic forgetting during single-pass online continual learning under convex objectives. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Investigates catastrophic forgetting bounds, not model compression or efficiency.
Groups Don't Move as One: A Bregman Clustering View of Linearized Group Influence This paper uses Bregman clustering to address opposing member effects within linearized group influence estimates for machine learning models. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Studies dataset attribution methods, not model compression or distillation.
Machine Learning-Based Prediction of Dengue Infection Using Clinical and Hematological Data from Bangladesh: A Comparative Analysis of Benchmark and Extended Prediction Models This paper evaluates standard machine learning models to predict dengue infection from clinical and hematological data in Bangladesh. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Studies dengue prediction models, not efficient model serving techniques.
Explore This! Beat and Downbeat Tracking From a Learned Tatum Grid This paper proposes a neural architecture that predicts a tatum grid to track musical beats and downbeats without dynamic Bayesian networks. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Focuses on musical beat tracking, not efficient language modeling.
Sequence-Structured Decision Learning for Branch-and-Bound in Mixed-Integer Linear Programming This paper introduces a sequence-structured framework that models the global search trajectory to learn branching and node selection in branch-and-bound. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Focuses on algorithmic search efficiency, not language model compression.
Physics-informed random neural architectures on compact boundaryless multiply connected manifolds for random-operator models in machine learning and uncertainty quantification This paper develops stochastic physics-informed neural networks on manifolds to solve stochastic boundary-value problems with random operators. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Deals with scientific computing architectures, not model compression or efficiency.
Disentangling Marginal Capacity from Geometry in Optimal-Transport Label Transfer This paper analyzes error patterns from class frequency imbalances in optimal-transport label transfer and proposes matching target composition marginals. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Focuses on distribution alignment in label transfer, not efficient model architectures or inference.
Terrarium: Growing Personal Assistants via Scalable Synthesis of App Worlds This paper presents a framework to synthesize executable application environments and training tasks from web information for personal language agents. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Focuses on synthesizing environment suites for assistants, not model compression or efficient inference.
World2TS: Real-World Event Context for Time-Series Forecasting This paper introduces a benchmark and dataset that pairs numerical time-series data with real-world event context for forecasting. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Focuses on time-series forecasting datasets, not small or efficient model architectures.
Blending experimental electron density information to protein-ligand interaction This paper introduces a pre-trained vision transformer that combines experimental electron-density maps with atomic coordinates to model protein-ligand interactions. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Studies multimodal molecular structural representations, not efficient LLM inference or compression.
Contrastive Linearized Laplace: Loss Geometry, Valid Repairs, and Reliable Cores This paper proposes a post-hoc framework based on Linearized Laplace approximation to quantify uncertainty across downstream tasks in contrastive learning. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Deals with post-hoc Laplace approximations, not model distillation, pruning, or efficiency.
Pro-t-VAE: Closed-Form Student's t Mixtures for Robust Prototype Discovery This paper introduces a variational autoencoder using a closed-form Student's t mixture prior for robust unsupervised image prototype discovery. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Studies Student's t mixture VAE formulations, not efficient LLM serving or compression.
FDTD Model of Noise Pollution from an AI Data Center This paper models low-frequency sound propagation over surrounding topography to evaluate noise pollution caused by AI data centers. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Models physical acoustic noise around data centers, not small or efficient AI architectures.
Prediction with Expert Advice under Heavy-Tailed Losses This paper establishes regret bounds and develops adaptive algorithms for prediction with expert advice under heavy-tailed losses. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Studies multi-armed bandit/expert advice theory, not model compression or efficient inference.
Minimum Volume Optimal Transport for Multivariate Conformal Prediction This paper presents a framework to minimize prediction-set volume in multivariate conformal prediction by learning a Mahalanobis ground cost. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Focuses on optimal transport conformal sets, not efficient model compression or serving.
TLA-Verus: Generating and Solving Temporal Proof Obligations with LLM Agents Uses language model agents and deterministic translation to convert TLA+ formal specifications into machine-checkable Verus proofs. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Focuses on formal software verification with LLMs, not model compression or efficient serving.
An Algebraic Approach to Constrained K-means Clustering Uses Gröbner bases to simplify pairwise constraints and improve assignment efficiency in constrained k-means clustering. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Studies k-means clustering constraints, not small or efficient language models.
CFG-Policy: Unified Diffusion Policy Learning Framework via Classifier-free Guidance Applies classifier-free guidance to diffusion policies to separate reinforcement learning exploration from core task capabilities. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Studies diffusion policy exploration in robotics, not efficient LLM compression or inference.
Compositional Average Velocity Modeling for Efficient Diffusion Offline Reinforcement Learning Proposes a single-step generative policy for offline reinforcement learning by modeling average velocity instead of iterative denoising. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Examines reinforcement learning diffusion sampling, not efficient LLM inference or distillation.
Exact Phase-Type Point Processes for Certified Measurement and Prediction of Event Streams Formulates a phase-type Hawkes process model with certified convex optimization to measure and predict financial event streams. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Addresses temporal point processes in finance, not efficient model architectures or compression.
Cross-Material Magnetic Core Loss Prediction via Operating-Condition Matching and Partial Optimal Transport Predicts magnetic core losses across materials using partial optimal transport to align operating conditions with limited data. Preprint	Small and efficient models via arxiv · 15 Oct, 06:18	Not relevant Applies optimal transport to magnetic loss, not efficient model compression or serving.

1,141–1,200 of 2,430← Previous Next →