Firdaus Janoos
Massachusetts Institute of Technology
38 Papers
206 Citations
Firdaus Janoos is an academic researcher from Massachusetts Institute of Technology. The author has contributed to research in topics: Computer science & Image segmentation. The author has an hindex of 15, co-authored 38 publications. Previous affiliations of Firdaus Janoos include ExxonMobil & Brigham and Women's Hospital.
Chat about Author
Papers
•Proceedings Article
Implementation Matters in Deep RL: A Case Study on PPO and TRPO
Logan Engstrom,Andrew Ilyas,Shibani Santurkar,Dimitris Tsipras,Firdaus Janoos,Larry Rudolph,Aleksander Madry +6 more
- 30 Apr 2020
TL;DR: The results show that algorithm augmentations found only in implementations or described as auxiliary details to the core algorithm are responsible for most of PPO's gain in cumulative reward over TRPO, and fundamentally change how RL methods function.
•Posted Content
Implementation Matters in Deep Policy Gradients: A Case Study on PPO and TRPO
Logan Engstrom,Andrew Ilyas,Shibani Santurkar,Dimitris Tsipras,Firdaus Janoos,Larry Rudolph,Aleksander Madry +6 more
TL;DR: The consequences of "code-level optimizations:" algorithm augmentations found only in implementations or described as auxiliary details to the core algorithm are investigated, and they are responsible for most of PPO's gain in cumulative reward over TRPO and fundamentally change how RL methods function.
198
Bayesian characterization of uncertainty in intra-subject non-rigid registration
TL;DR: A Bayesian non-rigid registration framework where conventional dissimilarity and regularization energies can be included in the likelihood and the prior distribution on deformations respectively through the use of Boltzmann's distribution is proposed.
76
Robust 3D reconstruction and identification of dendritic spines from optical microscopy imaging.
Firdaus Janoos,Kishore R. Mosaliganti,Xiaoyin Xu,Xiaoyin Xu,Raghu Machiraju,Kun Huang,Stephen T. C. Wong +6 more
TL;DR: The skeleton of the dendrite is extracted by a procedure based on the medial geodesic function that is robust and topology preserving, and it is used to accurately identify spines and the method is shown to be robust.
66
Patent
System and method for monitoring safety and productivity of physical tasks
Haytham Elhawary,Aditya Bansal,Firdaus Janoos,Selim Youssry +3 more
- 12 May 2017
TL;DR: In this article, a method for monitoring workplace safety and evaluating risks during lifting activities is presented, the method comprising receiving signals from wearable first and second devices, identifying portions of the signals corresponding to lifting activities.
33