• Publications
  • Influence
Soft Actor-Critic Algorithms and Applications
TLDR
Soft Actor-Critic (SAC), the recently introduced off-policy actor-critic algorithm based on the maximum entropy RL framework, achieves state-of-the-art performance, outperforming prior on-policy and off- policy methods in sample-efficiency and asymptotic performance.
Diversity is All You Need: Learning Skills without a Reward Function
TLDR
The proposed DIAYN ("Diversity is All You Need"), a method for learning useful skills without a reward function, learns skills by maximizing an information theoretic objective using a maximum entropy policy.
Multifactorial Evolution: Toward Evolutionary Multitasking
TLDR
This paper formalizes the concept of evolutionary multitasking and proposes an algorithm to handle multiple optimization problems simultaneously using a single population of evolving individuals and develops a cross-domain optimization platform that allows one to solve diverse problems concurrently.
Gradient Surgery for Multi-Task Learning
TLDR
This work identifies a set of three conditions of the multi-task optimization landscape that cause detrimental gradient interference, and develops a simple yet general approach for avoiding such interference between task gradients.
Learning Complex Dexterous Manipulation with Deep Reinforcement Learning and Demonstrations
TLDR
This work shows that model-free DRL with natural policy gradients can effectively scale up to complex manipulation tasks with a high-dimensional 24-DoF hand, and solve them from scratch in simulated experiments.
Imitation from Observation: Learning to Imitate Behaviors from Raw Video via Context Translation
TLDR
This work proposes an imitation learning method based on video prediction with context translation and deep reinforcement learning that enables a variety of interesting applications, including learning robotic skills that involve tool use simply by observing videos of human tool use.
Evolutionary Multitasking for Multiobjective Continuous Optimization: Benchmark Problems, Performance Metrics and Baseline Results
TLDR
Nine test problems for multi-task multi-Objective optimization (MTMOO), each of which consists of two multiobjective optimization tasks that need to be solved simultaneously, are suggested.
Learning Invariant Feature Spaces to Transfer Skills with Reinforcement Learning
TLDR
This paper introduces a problem formulation where two agents are tasked with learning multiple skills by sharing information and uses the skills that were learned by both agents to train invariant feature spaces that can be used to transfer other skills from one agent to another.
Multiobjective Multifactorial Optimization in Evolutionary Multitasking
TLDR
This paper presents a realization of the evolutionary multitasking paradigm within the domain of multiobjective optimization, which leads to the possibility of automated transfer of information across different optimization exercises that may share underlying similarities, thereby facilitating improved convergence characteristics.
On computing the data cube
TLDR
This paper focuses on a special case of the aggregation problem|computation of the cube operator, which requires computing group-bys on all possible combinations of a list of attributes, and extends hash-based and sort-based grouping methods with several optimizations.
...
1
2
3
4
5
...