Neural Diffeomorphic Non-uniform B-spline Flows

Authors: Seongmin Hong, Se Young Chun

AAAI 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Lastly, we performed experiments on solving the force matching problem in Boltzmann generators, demonstrating that our C2-diffeomorphic non-uniform B-spline flows yielded solutions better than previous spline flows and faster than smooth normalizing flows.
Researcher Affiliation Academia Seongmin Hong1, Se Young Chun1,2* 1 Department of Electrical and Computer Engineering, Seoul National University, Republic of Korea 2 INMC, Interdisciplinary Program in AI, Seoul National University, Republic of Korea
Pseudocode Yes Algorithm 1: Non-uniform B-spline parameter generation
Open Source Code Yes Our source code is publicly available at https://github.com/smhongok/Non-uniform-B-spline-Flow.
Open Datasets No The paper mentions using "alanine dipeptide" for experiments but does not provide concrete access information (link, DOI, repository, or formal citation with authors/year) for its public availability.
Dataset Splits No The paper mentions training and testing but does not provide specific details on dataset splits (e.g., percentages or sample counts for train, validation, and test sets) within the main text.
Hardware Specification Yes All computations were conducted on NVIDIA Ge Force RTX3090.
Software Dependencies No The paper does not provide specific software dependencies with version numbers (e.g., Python, PyTorch, CUDA versions).
Experiment Setup Yes Two loss functions were used: NLL loss LNLL or NLL + FM loss (1 λFM)LNLL + λFMLFM where f(x(n)) x log px(x(n)) 2 Ex p x h f(x) x log px(x) 2 2 i , and we set λFM = 0.001.