Compas
PulseAugur coverage of Compas — every cluster mentioning Compas across labs, papers, and developer communities, ranked by signal.
4 day(s) with sentiment data
-
New fairness framework analyzes utility functions, not just policies
Researchers have proposed a new framework for analyzing fairness in machine learning by focusing on the utility function itself, rather than imposing constraints on the predictive policy. This approach, termed 'value of…
-
New DECAF method ensures fairness across synthetic data generators
Researchers have developed a method called DECAF to ensure fairness in synthetic data, applicable across various data generation techniques including GANs and diffusion models. This approach was tested on the Adult and …
-
New DECAF framework ensures fairness across synthetic data generators
A new framework called DECAF has been developed to ensure fairness across various synthetic data generators. This framework allows statistical agencies and regulators to shape data generators to eliminate unfair pathway…
-
AI failures stem from unanswered ethical questions, not just engineering
Many AI system failures stem not from technical flaws but from unaddressed ethical and societal questions. Systems like COMPAS, Tay, Clearview AI, and the Robodebt scheme were deployed without adequately considering the…
-
COMPAS method optimizes code generation by jointly tuning models, prompts, and settings
Researchers have developed COMPAS, a novel method for optimizing code generation by jointly searching over models, prompts, and decoding settings. This difficulty-aware approach learns group-specific quality-cost fronts…
-
New framework fools AI explainability auditors by embedding evasion logic
Researchers have developed a new framework called "Crushing the Evidence" that can fool white-box explainable AI (XAI) auditors. This dual-penalty evasion technique embeds evasion logic directly into model parameters, a…
-
New FairDiffuseVQVAE model enhances synthetic data fairness without sacrificing quality
Researchers have developed FairDiffuseVQVAE, a novel two-stage architecture for generating synthetic tabular data that decouples data fidelity from fairness. The first stage uses a vector-quantized autoencoder for recon…
-
New CFQ method improves recourse stability in quantized AI models
Researchers have developed a new method called Counterfactual-Faithful Quantization (CFQ) to address issues with model quantization in decision systems that offer algorithmic recourse. Standard quantization can alter th…
-
New metric PCER audits fairness in differentially private ML
Researchers have introduced a new group fairness criterion called the Privacy-Cost Equity Ratio (PCER) for differentially private machine learning systems. PCER addresses the issue that differential privacy mechanisms l…
-
New framework characterizes utility-separation trade-off in ML models
Researchers have developed a new information-theoretic framework to characterize the trade-off between utility and separation in machine learning models. This framework proves the concavity of the utility-separation Par…
-
Study finds global LLM leaderboards misleading, proposes portfolio rankings
A new research paper argues that current leaderboards for large language models (LLMs) are misleading due to significant heterogeneity in user preferences across languages and tasks. The study analyzed approximately 89,…