/Reading_record

My reading papers about RL, AI

Reading_Record

It is my reading with RL, AI papers. There have 7 levels to explain the myself learning on reachers.

There are few class :

  1. Grasping

  2. Computer vision

  3. Transfering and Muilt-task learning

  4. Manipulation

  5. Exploration

  6. Sim2real

  7. Others

Grasping

paper ~15% ~30% ~45% ~60% ~70% ~80% ~90%
Deep Reinforcement Learning for Vision-Based Robotic Grasping: A Simulated Comparative Evaluation of Off-Policy Methods ✓ ✓ ✓ ✓ ✓ ✓
Learning Hand-Eye Coordination for Robotic Grasping with Deep Learning and Large-Scale Data Collection ✓ ✓ ✓ ✓ ✓ ✓
A Framework for Efficient Robotic Manipulation ✓ ✓
Grasp Proposal Networks- An End-to-End Solution for Visual Learning of Robotic Grasps ✓ ✓ ✓
QT-Opt Scalable Deep_Reinforcement Learning for Vision-Based Robotic Manipulation ✓ ✓ ✓ ✓ ✓
Dex-Net 2.0: Deep Learning to Plan Robust Grasps withSynthetic Point Clouds and Analytic Grasp Metrics ✓ ✓ ✓ ✓ ✓ ✓
Dex-Net 2.1: Learning Deep Policies for Robot Bin Picking by Simulating Robust Grasping Sequences ✓ ✓ ✓ ✓
Dex-Net 3.0: Computing Robust Robot Suction Grasp Targets using a New Analytic Model and Deep Learning ✓ ✓ ✓ ✓
Combining Deep Deterministic Policy Gradient with Cross-Entropy Method ✓ ✓
Learning_Hand-Eye_Coordination_for_Robotic_Grasping_with_Deep_Learning_and_Large-Scale_Data_Collection ✓ ✓ ✓ ✓

Computer vision

paper ~15% ~30% ~45% ~60% ~70% ~80% ~90%
You Only Look Once: Unified, Real-Time Object Detection ✓ ✓ ✓ ✓ ✓ ✓ ✓
YOLO9000: Better, Faster, Stronger ✓ ✓ ✓ ✓ ✓ ✓ ✓
YOLOv3: An Incremental Improvement ✓ ✓ ✓ ✓ ✓ ✓
Mask R-CNN ✓ ✓ ✓ ✓ ✓ ✓

Transfering and Muilt-task learning

paper ~15% ~30% ~45% ~60% ~70% ~80% ~90%
EPOpt: Learning robust neural network policies.. ✓ ✓
Sim-to-Real Transfer of Robotic Control with Dynamics Randomization ✓
Adapting Visuomotor Representations with Weak Pairwise Constraints ✓

Manipulation

paper ~15% ~30% ~45% ~60% ~70% ~80% ~90%
Transporter Networks- Rearranging the Visual World for Robotic Manipulation ✓ ✓ ✓ ✓
Robotic Table Tennis with Model-Free Reinforcement Learning ✓ ✓

Exploration

paper ~15% ~30% ~45% ~60% ~70% ~80% ~90%
Deep Exploration via Bootstrapped DQN ✓ ✓ ✓
VIME_Variational Information Maximizing Exploration ✓ ✓

Sim2real

paper ~15% ~30% ~45% ~60% ~70% ~80% ~90%
Using Simulation and Domain Adaptation to Improve Efficiency of Deep Robotic Grasping ✓ ✓ ✓
3D Simulation for Robot Arm Control with Deep Q-Learning ✓ ✓ ✓ ✓ ✓
RL-CycleGAN Reinforcement Learning_Aware Simulation_To_Real ✓ ✓ ✓ ✓ ✓ ✓
RetinaGAN An Object-aware Approach to Sim-to-Real Transfer ✓ ✓ ✓ ✓ ✓
Sim-To-Real Transfer for Miniature AutonomousCar Racing ✓ ✓ ✓

Others

paper ~15% ~30% ~45% ~60% ~70% ~80% ~90%
Learning to Stop: Dynamic Simulation Monte-Carlo Tree Search ✓
An End-to-End Trainable Neural Network for Image-based Sequence Recognition and Its Application to Scene Text Recognition ✓ ✓ ✓
ClearGrasp 3D Shape Estimation of Transparent Objects for Manipulation ✓ ✓ ✓