Markov chain
PulseAugur coverage of Markov chain — every cluster mentioning Markov chain across labs, papers, and developer communities, ranked by signal.
8 day(s) with sentiment data
-
New statistical kernels analyzed for Markov chains
Researchers have developed new theoretical tools to analyze the statistical properties of sliding-window count kernels derived from stationary Markov chains. The study establishes spectral-gap bounds and Poincaré inequa…
-
New research offers faster Markov chain convergence methods
Two new research papers propose novel methods for accelerating Markov chain convergence. The first paper introduces a criterion called asymptotic equivalence with the target, offering a direct route to convergence proof…
-
Gaussian Perturbations Prevent Oversmoothing in Recurrent GNNs
Researchers have developed a novel method using persistent Gaussian perturbations to combat oversmoothing in recurrent graph neural networks (GNNs). This technique injects independent Gaussian noise after each propagati…
-
Two arXiv papers detail learning dynamical systems from single trajectories · 2 sources tracked
Two new research papers submitted to arXiv's stat.ML section explore the learning of dynamical systems from single trajectories. The first paper focuses on switched non-linear dynamical systems, providing theoretical gu…
-
LLM text processing explained: from word counts to linguistics and semiotics · 8 sources tracked
A series of articles explores the technical underpinnings of how Large Language Models (LLMs) process and understand text. The author delves into various methods, from basic word counting and statistical techniques like…
-
New framework classifies Thompson Sampling under model misspecification
This paper introduces a novel stochastic stability framework to analyze Thompson Sampling (TS) algorithms in dynamic decision-making scenarios where the underlying model might be misspecified. The research provides a de…
-
New framework interprets quantum learning models via stochastic processes
A new research paper proposes a framework to interpret quantum learning models by representing them as stochastic processes. The work, led by Johannes Fankhauser, addresses the challenge that quantum dynamics typically …
-
Researchers analyze Transformer attention mechanisms and feed-forward networks · 2 sources tracked
Two new research papers explore the fundamental components of Transformer models, specifically focusing on the role of attention mechanisms versus feed-forward networks. The first paper, "A Controlled Study of Attention…
-
New diffusion distance metric measures spatial clustering beyond local patterns
Researchers have introduced a new metric called diffusion distance to measure spatial clustering. This metric extends traditional spatial autocorrelation measures like Moran's I by considering global graph geometry rath…
-
MAESTRO framework improves MoE model pruning by modeling expert dependencies
Researchers have developed MAESTRO, a novel structured pruning framework designed to address the deployment bottleneck in Mixture-of-Experts (MoE) language models. Unlike previous methods that use local heuristics, MAES…
-
AI predicts 5G network states to overcome backhaul delay
Researchers have developed a novel two-stage predictive framework to mitigate the impact of backhaul delay in coordinated beamforming for 5G networks. The framework utilizes a Spectral Temporal Graph Neural Network (Ste…
-
New fuzzy logic framework enhances group decision-making with dynamic weighting
This paper introduces a new framework called the Double Fuzzy Probabilistic Interval Linguistic Term Set (DFPILTS) to address limitations in existing probabilistic linguistic term methods for group decision-making. The …
-
Transformers learn sparse attention patterns incrementally, study finds
This paper investigates how transformers learn sparse attention patterns incrementally when trained on a high-order Markov chain. Researchers Oğuz Kaan Yüksel and colleagues observed that transformers learn by first foc…
-
New SAGO framework enables real-time 3D Gaussian segmentation
Researchers have developed SAGO (Segment Any Gaussians Online), a new framework designed to enable real-time interactive segmentation of 3D Gaussian Splatting (3DGS) scenes. Unlike previous methods that required lengthy…
-
Random Forest Ensemble Size Tuning Explained by New Stationary Distribution Theory · 2 sources tracked
This paper introduces a theoretical framework for understanding the stationary distribution of ensemble sizes in Random Forests during plateau-based tuning. The research models the central ensemble size as a birth-death…
-
New method optimizes PL-SGD with Markovian noise for improved bounds
Researchers have developed a new method for optimizing smooth objectives that satisfy the Polyak-Łojasiewicz (PL) condition, particularly when gradient samples are influenced by Markovian noise. This approach establishe…
-
New TD(0) algorithm achieves robust and fast convergence with single stepsize
Researchers have developed a new method for linear TD(0) algorithms that uses a single stepsize schedule, eliminating the need for prior knowledge of curvature parameters. This approach provides high-probability guarant…
-
New TR-CIE sampler enhances discrete flow matching quality with limited function evaluations · 3 sources tracked
Researchers have developed a new sampling method called the Time-Reparameterized Cumulative Intensity Extrapolation (TR-CIE) sampler for discrete flow matching (DFM). This method aims to enhance sampling quality in gene…
-
New Doeblin Curves Offer Finer-Grained Contraction Guarantees
Researchers have introduced the concept of a "Doeblin curve" to provide a more detailed characterization of multi-way contraction behavior in Markov kernels. This new approach offers non-vacuous contraction guarantees e…
-
New DRRL Algorithm Achieves Finite-Time Convergence with Linear Approximation
Researchers have developed a new algorithm for Distributionally Robust Reinforcement Learning (DRRL) that provides finite-time convergence guarantees even with linear function approximation. This algorithm addresses lim…