Data Shapley
PulseAugur coverage of Data Shapley — every cluster mentioning Data Shapley across labs, papers, and developer communities, ranked by signal.
1 day(s) with sentiment data
-
New research questions per-point data valuation for AI training
A new research paper argues that common methods for selecting training data, such as using per-point data-valuation scores like Data Shapley, are flawed. The paper explains that a data point's value is not fixed but dep…
-
New framework quantifies marginal data value for enterprises
Researchers have introduced the Data Differentiator Framework (DDF), a novel method for evaluating candidate datasets. DDF combines a redundancy penalty based on mutual information with a cost-normalized marginal utilit…
-
New system enables private and verifiable speech-agent marketplaces
Researchers have developed WaveCommitments, a novel system designed for private and verifiable marketplaces where speech agents can trade assets like embeddings or raw data. This system utilizes multivariate polynomial …
-
New algorithms offer exact Data Shapley for weighted KNN regression
Researchers have developed new algorithms for calculating Data Shapley values in weighted k-nearest-neighbor (KNN) regression and soft-label prediction. Previously, exact calculations for these scenarios were computatio…
-
New framework unifies data selection via sequential decision-making
Researchers have developed a new framework that treats data selection as a sequential decision-making problem, utilizing dynamic programming to find optimal selection sequences. This approach unifies existing methods li…
-
NASH framework enhances data selection for machine learning
Researchers have introduced NASH, a new framework for data selection in machine learning that aims to improve the effectiveness of methods like Data Shapley. NASH decomposes utility functions into simpler, Shapley-infor…
-
New methods offer efficient data valuation for LLMs and VLMs
Two new research papers propose novel methods for data valuation in large language models (LLMs). The first, "For-Value," introduces an efficient forward-only framework that estimates data value using a single forward p…