One-two page summary/review of research papers read by me in Reinforcement Learning along with possible improvements which I could think of.
Paper | Authors | Conference | Authors Affiliation |
---|---|---|---|
Object Sensitive Deep Reinforcement Learning | Y.Li, K.Sycara, R.Iyer | GCAI-17 | Carnegie Mellon University |
Safe Reinforcement Learning with Model Uncertainty Estimates | B.Lutjens, M.Everett, J.How | ICRA-18 | Massachusetts Institute of Technology |
Multi-stage Reinforcement Learning for Object Detection | J.Konig, S.Malberg, M.Martens, S.Niehaus, A.Grimberghe, A.Ramaswamy | CVC-19 | Paderborn University |
Curiosity-driven Exploration by Self-supervised Prediction | D.Pathak, P.Agrawal, A.Efros, T.Darrell | ICML-17 | University of California Berkeley |
AMC: AutoML for Model Compression and Acceleration on Mobile Devices | Y.He, J.Lin, Z.Liu, H.Wang, L.Li, S.Han | ECCV-18 | Massachusetts Institute of Technology, Carnegie Mellon University |
Asynchronous Methods for Deep Reinforcement Learning | V.Mnih, A.Badia, M.Mirza, A.Graves, T.Harley, T.Lillicrap, D.Silver, K.Kavukcuoglu | ICML-16 | Google DeepMind, Montreal Institute of Learning Algorithms |
Robust Adversarial Reinforcement Learning | L.Pinto, J.Davidson, R.Sukthankar, A.Gupta | ICML-17 | Carnegie Mellon University, Google Brain, Google Research |
- Writing paper summaries/reviews does help me in retaining the information in the paper and understanding it in a better way as it makes me think more about the paper which I would not have done by just reading the papers.
- It also helps me to think of possible extensions of the project by connecting information in the previously read papers and trying to combine the positives of all the papers.
- Motivates me to read more papers.
The paper review/summary contains ideas which I could think of when I read the paper. The ideas/extensions may or may not work. They are just to make me think more about the paper. I do not intend to insult or degrade any of the authors' work.