Communications of the ACM
Latest Research
From Communications of the ACM
Technical Perspective: Breaking the Mold of Machine Learning
"Never-Ending Learning" is the latest and one of the most compelling incarnations of Tom Mitchell and his collaborators' research investigating how to broaden the...
From Communications of the ACM
Never-Ending Learning
In this paper we define more precisely the never-ending learning paradigm for machine learning, and present one case study: the Never-Ending Language Learner (NELL)...
From Communications of the ACM
Technical Perspective: Expressive Probabilistic Models and Scalable Method of Moments
The authors of "Learning Topic Models—Provably and Efficiently," developed a new method for fitting topic models and at large scale.
From Communications of the ACM
Learning Topic Models - Provably and Efficiently
This article shows that some new theoretical algorithms that have provable guarantees can be adapted to yield highly practical tools for topic modeling.
From Communications of the ACM
Technical Perspective: Can High Performance Be Portable?
"Halide: Decoupling Algorithms from Schedules for High-Performance Image Processing" by Ragan-Kelley et al. on the image processing language Halide explores a substantially...
From Communications of the ACM
Halide: Decoupling Algorithms from Schedules For High-Performance Image Processing
We propose a new programming language for image processing pipelines, called Halide, that separates the algorithm from its schedule.
From Communications of the ACM
Technical Perspective: Exploring a Kingdom By Geodesic Measures
"The Heat Method for Distance Computation," by Crane, Weischedel, and Wardetzky, shows that the gradient of the probability density function of a random walk is...
From Communications of the ACM
The Heat Method For Distance Computation
We introduce the heat method for solving the single- or multiple-source shortest path problem on both flat and curved domains.
From Communications of the ACM
Technical Perspective: Linking Form, Function, and Fabrication
To avoid costly feedback loops between design, engineering, and fabrication, research in computer graphics has recently tried to incorporate key aspects of function...
From Communications of the ACM
Spin-It: Optimizing Moment of Inertia For Spinnable Objects
In this article, we describe an algorithm to generate designs for spinning objects by optimizing their mass distribution.
From Communications of the ACM
Technical Perspective: Unexpected Connections
The inherent scalability of an interface is the focus of "The Scalable Commutativity Rule" by Austin T. Clements, et al.
From Communications of the ACM
The Scalable Commutativity Rule: Designing Scalable Software For Multicore Processors
This paper introduces an interface-driven approach to building scalable software.
From Communications of the ACM
Technical Perspective: Ironfleet Simplifies Proving Safety and Liveness Properties
"IronFleet: Proving Safety and Liveness of Practical Distributed Systems," by Chris Hawblitzel, et al., describes mechanically checked proofs for two non-trivial...
From Communications of the ACM
Ironfleet: Proving Safety and Liveness of Practical Distributed Systems
We demonstrate the methodology on a complex implementation of a Paxos-based replicated state machine library and a lease-based sharded key-value store. With our...
From Communications of the ACM
Technical Perspective: Low-Depth Arithmetic Circuits
The past few years have seen a revolution in our understanding of arithmetic circuits. "Unexpected Power of Low-Depth Arithmetic Circuits" by Gupta et al. on the...
From Communications of the ACM
Unexpected Power of Low-Depth Arithmetic Circuits
Several earlier results have shown that it is possible to rearrange basic computational elements in surprising ways to give more efficient algorithms. The main...
From Communications of the ACM
Technical Perspective: What Led Computer Vision to Deep Learning?
We are in the middle of the third wave of interest in artificial neural networks as the leading paradigm for machine learning. "ImageNet Classification with Deep...
From Communications of the ACM
ImageNet Classification with Deep Convolutional Neural Networks
In the 1980s backpropagation did not live up to the very high expectations of its advocates. Twenty years later, we know what went wrong: for deep neural networks...
From Communications of the ACM
DeepDive: Declarative Knowledge Base Construction
We describe DeepDive, a system that combines database and machine learning ideas to help to develop knowledge base construction systems.
From Communications of the ACM