Shapley value
PulseAugur coverage of Shapley value — every cluster mentioning Shapley value across labs, papers, and developer communities, ranked by signal.
6 day(s) with sentiment data
-
New game theory approach offers stability bands for feature attribution
Researchers have introduced a new framework for understanding feature attributions in machine learning models, moving beyond single-value explanations to incorporate coalition dynamics. This approach utilizes concepts f…
-
New framework integrates fairness and explainability using Shapley values
A new framework integrates fairness and explainability in machine learning by utilizing Shapley values. This approach allows for the measurement and explanation of unfairness, identifying features that contribute to it.…
-
New method improves Shapley value computation for AI explainability
Researchers have developed a new sampling-based method to more efficiently calculate Shapley values, a metric used for explaining AI models. This approach leverages Owen's multilinear extension of games and combines str…
-
New method LEDGER estimates memory value in retrieval systems
A new paper introduces LEDGER, a method designed to address a limitation in top-k retrieval systems where the exact value of individual memories cannot be determined from retrieval logs alone. The research highlights th…
-
New method improves feature attribution for computer vision models
Researchers have developed a new method for feature attribution in machine learning, specifically addressing challenges in computer vision where features can act as contextual distractors. The proposed approach, based o…
-
Neural Tangent Kernel Analysis Sheds Light on PDF Training Dynamics
Researchers have analyzed the training dynamics of feed-forward neural networks using a Neural Tangent Kernel (NTK) approach applied to Parton Distribution Functions (PDFs) derived from deep inelastic scattering (DIS) d…
-
New Spectral-Spatial Interpretation Method Enhances AI Model Explanations
Researchers have introduced Spectral-Spatial Interpretation (SSI), a novel method for explaining AI model behavior on multispectral image data. Unlike traditional methods that provide separate spatial heatmaps or global…
-
New diffusion model optimizes urban flow monitoring and sensor placement
Researchers have developed Diff-SPORT, a novel diffusion-based framework designed for optimizing sensor placement and reconstructing turbulent flows in urban environments. This method utilizes a generative diffusion pri…
-
New research tackles Shapley value estimation challenges in ML · 2 papers
Two new research papers published on arXiv propose novel methods for estimating Shapley values, a key technique in machine learning for feature attribution. The first paper introduces FUSHAP, designed to handle multi-si…
-
New 'Local Shapley' method drastically cuts data valuation computation
Researchers have introduced "Local Shapley," a novel method for data valuation that significantly reduces computational complexity. Unlike traditional approaches that consider all possible training data combinations, Lo…
-
New framework uses Shapley values for data privacy and utility
Researchers have developed a new framework that adapts Shapley-value-based feature attribution to address data privacy concerns. This approach captures both disclosure risk and data utility at the feature level, offerin…
-
New method accurately attributes data contributions for Adam optimizer
Researchers have developed a new method called Adam-Aware In-Run Data Shapley to accurately attribute data contributions in machine learning models that use the Adam optimizer. Existing methods, which rely on the linear…
-
New nonlinear methods improve AI attribution over Shapley value
Researchers have developed new nonlinear axiomatic attribution methods to address limitations in the traditional Shapley value, particularly its linearity which can obscure important player contributions. These novel me…
-
New framework proposes fair compensation for LLM-summarized content
A new research paper proposes a framework using Shapley values to address fair compensation for content creators whose work is summarized by large language models (LLMs). The proposed method, called Cluster Shapley, app…
-
Shapley values enhance LLM explainability in finance
Researchers have developed a new method using Shapley values to explain the behavior of large language models (LLMs) in financial applications. This approach aims to align LLM explanations with established financial dom…
-
New method explains outliers in interval-valued data using Shapley values
Researchers have developed a new method for explaining outliers in interval-valued data, addressing a gap in current outlier detection techniques. This approach utilizes Shapley values to provide a detailed breakdown of…
-
New DAG-SHAP method improves feature attribution in causal AI models
Researchers have introduced DAG-SHAP, a novel feature attribution method designed for directed acyclic graphs (DAGs) that addresses limitations of existing Shapley value-based approaches. Unlike previous node-centric me…
-
ShapKAN framework enhances KAN interpretability and compression
A new framework called ShapKAN has been developed to address the challenges of pruning Kolmogorov-Arnold Networks (KANs). This method utilizes Shapley values to evaluate node importance in a manner that is invariant to …
-
New framework attributes data contributor value in text-to-image models
Researchers have developed SurrogateSHAP, a novel framework designed to efficiently attribute contributions to data contributors in text-to-image models. This method avoids the computationally expensive process of retra…
-
New AI Framework Simplifies Patent Valuation
Researchers have developed PatentXAI, a novel framework designed to tackle the complex problem of valuing individual patents within large product portfolios. This framework leverages graph-conditioned hierarchical Shapl…