AlphaLab automates end-to-end scientific experimentation using model-generated adapters and adversarial evaluation, achieving significant performance gains in CUDA optimization and machine learning.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
AlphaLab: Autonomous Multi-Agent Research Across Optimization Domains with Frontier LLMs
AlphaLab automates end-to-end scientific experimentation using model-generated adapters and adversarial evaluation, achieving significant performance gains in CUDA optimization and machine learning.