Jensen-Shannon divergence
PulseAugur coverage of Jensen-Shannon divergence — every cluster mentioning Jensen-Shannon divergence across labs, papers, and developer communities, ranked by signal.
5 day(s) with sentiment data
-
LLMs generate realistic travel diaries, outperforming traditional methods in purpose and mode prediction
Researchers have developed a novel method using Large Language Models (LLMs) to generate individual travel diaries for transportation modeling. This approach synthesizes personas from open-source data like the American …
-
CyclOT framework learns quadratic optimal transport maps from unpaired data
Researchers have introduced CyclOT, a novel neural framework for learning quadratic optimal transport maps from unpaired samples in high dimensions. This bidirectional approach utilizes synchronized forward-backward int…
-
New Bayesian method improves multi-agent AI decision-making
Researchers have developed a novel approach to multi-agent collective decision-making using Bayesian backward reasoning. This method aims to improve performance when multiple AI agents produce conflicting answers by ana…
-
New benchmarks and frameworks enhance multimodal AI reasoning and reliability
Researchers have developed new benchmarks and frameworks to improve the reliability and evidence-grounded reasoning of multimodal AI agents. Sci-MMR, a benchmark for scientific reasoning, highlights a significant gap be…
-
Neural audio codec tokens show language-like statistical properties
A new paper analyzes the statistical properties of tokens generated by neural audio codecs, finding that these sequences exhibit language-like characteristics. The study evaluated 13 different neural audio codecs across…
-
New Markov model technique analyzes musical structure in Debussy's "Syrinx"
Researchers have developed a novel method for estimating transition kernels in Markov models, utilizing overlapping sliding windows to derive local kernels directly from symbolic sequences. This observation-driven appro…
-
Diffusion Language Models Advance with New Efficiency and Safety Techniques · 10 sources tracked
Recent research explores advancements in diffusion language models (DLMs), focusing on improving their efficiency, safety, and capabilities. Papers introduce methods like Q-Skew for privacy risk assessment and PII extra…
-
New metric quantifies explanation consistency in medical AI fairness
Researchers have introduced a new metric called the Explanation Consistency Score (ECS) to evaluate fairness in medical imaging models. This score, based on Jensen-Shannon divergence, quantifies how similar the attribut…
-
CrevasseSeg framework uses label-efficient methods for UAV glacier mapping
Researchers have developed CrevasseSeg, a framework designed for efficient segmentation of glacier crevasses using uncrewed aerial vehicle (UAV) imagery. This approach aims to reduce the need for extensive pixel-level a…
-
LSTM networks enhance electricity price prediction with adaptive learning
Researchers have developed an adaptive online learning framework using Long Short-Term Memory (LSTM) networks to improve the accuracy of day-ahead electricity price predictions in the California energy market. The model…
-
New theory enables modular training of robust large language models
Researchers have developed a theoretical framework for modularly training large language models (LLMs) by combining smaller, domain-specific expert models. This approach aims to achieve performance comparable to monolit…
-
AI maps lung cancer growth patterns using visual vocabulary
Researchers have developed a novel weakly supervised Bag-of-Visual-Words (BoVW) pipeline to map lung adenocarcinoma growth patterns from whole slide images. This method utilizes frozen foundation model embeddings to lea…
-
New HEIMAT framework automatically debiases language models
Researchers have developed a new framework called HEIMAT to automatically debias language models. This framework addresses limitations of existing methods, such as high computational costs, scalability issues, and the n…
-
Divergence Decoding fuses LLM capabilities without retraining
Researchers have introduced Divergence Decoding, a novel training-free framework designed to fuse the capabilities of specialized scientific language models with generalist models. This method uses Jensen-Shannon diverg…
-
New wavelet denoising framework targets network anomaly detection
Researchers have developed a novel drift-aware framework for wavelet denoising specifically designed for network traffic anomaly detection. This approach treats adaptive wavelet denoising as a preprocessing step optimiz…
-
New method Soft Clamp combats AI agent over-calling of tools
Researchers have identified a failure mode in multi-teacher on-policy distillation for AI agents that use tools. This method, while improving tool-call recall, can cause agents to over-call tools inappropriately. The pa…
-
New research analyzes divergence measures for credit risk model monitoring
A new research paper published on arXiv analyzes the statistical properties and power of divergence measures for monitoring credit risk models. The study focuses on Jensen-Shannon Divergence and Kullback-Leibler Diverge…
-
New GAN Architecture SuRGe Enhances Image Super-Resolution
Researchers have developed Super-Resolution Generator (SuRGe), a novel Generative Adversarial Network (GAN) architecture designed to enhance image quality. SuRGe combines features from different network depths using lea…
-
New IHDec method secures LLM instruction hierarchies without fine-tuning · 2 sources tracked
Researchers have developed IHDec, a novel method to address instruction hierarchy failures in large language models (LLMs) during multi-turn conversations. Unlike previous solutions that require costly fine-tuning, IHDe…
-
SARA framework enhances multilingual capabilities in Mixture-of-Experts models
Researchers have introduced SARA (Semantically Anchored Routing Alignment), a new framework designed to improve the performance of Mixture-of-Experts (MoE) models in low-resource languages. SARA addresses the issue wher…