Caliban: Docker-based job manager for reproducible workflows
By S. Ritchie, Ambrose Slone, Vinay Venkatesh Ramasesh. 2020
7 citations recorded by the source corpus.
Moving between these environments almost always requires a nontrivial amount of effort on the part of the researcher, unrelated to the original research question. In Python, a common language for machine learning research, the researcher installs their script’s dependencies in a system-wide package registry. The Cloud environment can have different dependencies available, or different versions of the same dependency; different versions of Python itself; or different software drivers, which elicit different behavior from the script that seemed to work locally.
Subfields: Computer Networks and Communications, Information Systems, Information Systems and Management
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Computer Networks and Communications
In Information Systems
Citations in the corpus
- Anatomy of Catastrophic Forgetting: Hidden Representations and Task Semantics (2020, background): “[42] Sam Ritchie, Ambrose Slone, and Vinay Ramasesh.”
- Explaining neural scaling laws (2021, background): “Another direction for future research is to more explicitly derive within the theory the effects of “feature learning.””
- The Role of Permutation Invariance in Linear Mode Connectivity of Neural Networks (2021, methodology): “A IMPLEMENTATION DETAILS We used Caliban (Ritchie et al., 2020) to manage all experiments in a reproducible environment in Google Cloud’s AI Platform.”