Robert Nishihara
YOU?
Author Swipe
View article: ESCHER
ESCHER Open
As distributed applications become increasingly complex, so do their scheduling requirements. This development calls for cluster schedulers that are not only general, but also evolvable. Unfortunately, most existing cluster schedulers are …
View article: Hoplite
Hoplite Open
Task-based distributed frameworks (e.g., Ray, Dask, Hydro) have become increasingly popular for distributed applications that contain asynchronous and dynamic workloads, including asynchronous gradient descent, reinforcement learning, and …
View article: Hoplite: Efficient Collective Communication for Task-Based Distributed Systems.
Hoplite: Efficient Collective Communication for Task-Based Distributed Systems. Open
Collective communication systems such as MPI offer high performance group communication primitives at the cost of application flexibility. Today, an increasing number of distributed applications (e.g, reinforcement learning) require flexib…
View article: Hoplite: Efficient and Fault-Tolerant Collective Communication for Task-Based Distributed Systems
Hoplite: Efficient and Fault-Tolerant Collective Communication for Task-Based Distributed Systems Open
Task-based distributed frameworks (e.g., Ray, Dask, Hydro) have become increasingly popular for distributed applications that contain asynchronous and dynamic workloads, including asynchronous gradient descent, reinforcement learning, and …
View article: Lineage stash
Lineage stash Open
As cluster computing frameworks such as Spark, Dryad, Flink, and Ray are being deployed in mission critical applications and on larger and larger clusters, their ability to tolerate failures is growing in importance. These frameworks emplo…
View article: Policy Gradient Search: Online Planning and Expert Iteration without\n Search Trees
Policy Gradient Search: Online Planning and Expert Iteration without\n Search Trees Open
Monte Carlo Tree Search (MCTS) algorithms perform simulation-based search to\nimprove policies online. During search, the simulation policy is adapted to\nexplore the most promising lines of play. MCTS has been used by\nstate-of-the-art pr…
View article: Policy Gradient Search: Online Planning and Expert Iteration without Search Trees
Policy Gradient Search: Online Planning and Expert Iteration without Search Trees Open
Monte Carlo Tree Search (MCTS) algorithms perform simulation-based search to improve policies online. During search, the simulation policy is adapted to explore the most promising lines of play. MCTS has been used by state-of-the-art progr…
View article: On Systems and Algorithms for Distributed Machine Learning
On Systems and Algorithms for Distributed Machine Learning Open
The advent of algorithms capable of leveraging vast quantities of data and computational resources has led to the proliferation of systems and tools aimed to facilitate the development and usage of these algorithms. Hardware trends, includ…
View article: Tune: A Research Platform for Distributed Model Selection and Training
Tune: A Research Platform for Distributed Model Selection and Training Open
Modern machine learning algorithms are increasingly computationally demanding, requiring specialized hardware and distributed computation to achieve high performance in a reasonable time frame. Many hyperparameter search algorithms have be…
View article: Ray RLlib: A Framework for Distributed Reinforcement Learning
Ray RLlib: A Framework for Distributed Reinforcement Learning Open
Reinforcement learning (RL) algorithms involve the deep nesting of highly irregular computation patterns, each of which typically exhibits opportunities for distributed computation. We argue for distributing RL components in a composable w…
View article: RLlib: Abstractions for Distributed Reinforcement Learning
RLlib: Abstractions for Distributed Reinforcement Learning Open
Reinforcement learning (RL) algorithms involve the deep nesting of highly irregular computation patterns, each of which typically exhibits opportunities for distributed computation. We argue for distributing RL components in a composable w…
View article: Ray RLLib: A Composable and Scalable Reinforcement Learning Library
Ray RLLib: A Composable and Scalable Reinforcement Learning Library Open
Reinforcement learning (RL) algorithms involve the deep nesting of distinct components, where each component typically exhibits opportunities for distributed computation. Current RL libraries offer parallelism at the level of the entire pr…
View article: Ray: A Distributed Framework for Emerging AI Applications
Ray: A Distributed Framework for Emerging AI Applications Open
The next generation of AI applications will continuously interact with the environment and learn from these interactions. These applications impose new and demanding systems requirements, both in terms of performance and flexibility. In th…
View article: Discovering Causal Signals in Images
Discovering Causal Signals in Images Open
This paper establishes the existence of observable footprints that reveal the "causal dispositions" of the object categories appearing in collections of images. We achieve this goal in two steps. First, we take a learning approach to obser…
View article: Real-Time Machine Learning: The Missing Pieces
Real-Time Machine Learning: The Missing Pieces Open
Machine learning applications are increasingly deployed not only to serve predictions using static models, but also as tightly-integrated components of feedback loops involving dynamic, real-time decision making. These applications pose a …