Communications of the ACM
Multimodal AI Connects the Digital Dots
By knitting together multiple components and data streams, multimodal AI offers the promise of smarter, more human-like systems… From ACM NewsAn edited collection of advanced computing news from Communications of the ACM, ACM TechNews, other ACM resources, and news sites around the Web.
From ACM News
Breaking the Scaling Limits of Analog Computing
New technique could diminish errors that hamper the performance of super-fast analog optical neural networks.
From ACM TechNews
Study Urges Caution When Comparing Neural Networks to the Brain
A study by Massachusetts Institute of Technology researchers revealed the difficulty of comparing neural networks to the brain.
From ACM TechNews
Using AI to Control Digital Manufacturing
An international team of researchers taught a machine learning system with computer vision to monitor and control three-dimensional printing in real time.
From ACM TechNews
MIT Engineers Devise Recipe for Improving Any Autonomous Robotic System
A general-purpose optimizer aims to speed development of autonomous robotic systems by using an optimization code that automatically determines how to boost performance...
From ACM TechNews
Stronger Security for Smart Devices
Two new security techniques block power and electromagnetic side-channel attacks targeting analog-to-digital converters in smart devices.
From ACM TechNews
Hallucinating to Better Text Translation
Scientists developed the VALHALLA machine learning method to hallucinate images of written words, and using those to help translate that text into target languages...
From ACM TechNews
Unpacking Black-Box Models
A mathematical framework aims to quantify and evaluate the understandability of a machine learning model's explanations for its predictions.
From ACM TechNews
Engineers Use AI to Capture Complexity of Breaking Waves
An international research team lleveraged machine learning to model how waves break, which could help improve the design of offshore structures and climate prediction...
From ACM TechNews
An Easier Way to Teach Robots New Skills
Massachusetts Institute of Technology researchers have developed a machine learning method that enables robots to acquire new skills using a handful of human examples...
From ACM TechNews
Fighting Discrimination in Mortgage Lending
To help combat discrimination in mortgage lending, a new process can remove bias from the data used to train machine learning models.
From ACM TechNews
Injecting Fairness into ML Models
Researchers have devleoped a method of incorporating fairness into machine learning models, even if they are trained on unfair data.
From ACM TechNews
A Security Technique to Fool Would-Be Cyberattackers
A new technique to prevent memory timing side-channel attacks by malware exploits memory hardware common to multiple programs running on the same computer.
From ACM TechNews
AI System Rapidly Predicts How Two Proteins Will Attach
An artificial intelligence model can forecast the binding configuration of two proteins.
From ACM News
A New Language for Quantum Computing
Twist is an MIT-developed programming language that can describe and verify which pieces of data are entangled to prevent bugs in a quantum program.
From ACM TechNews
Generating a Realistic 3D World
Researchers developed the ThreeDWorld platform to simulate high-fidelity audio and visual environments.
From ACM TechNews
Taking Some Guesswork Out of Drug Discovery
The deep learning GeoMol model can rapidly predict the three-dimensional shapes of drug-like molecules, which could expedite drug discovery.
From ACM TechNews
Timber or Steel? Study Helps Builders Reduce Carbon Footprint of Truss Structures
Architects and engineers could design truss structures minimizing embodied carbon by using a new set of computational tools.
From ACM TechNews
AI Sheds Light on How the Brain Processes Language
Research by neuroscientists suggests the latest predictive language models' underlying mechanism functions similarly to the human brain's language-processing centers...
From ACM TechNews
Tool for Predicting Pedestrian Flow Expands Its Reach
Massachusetts Institute of Technology researchers have developed and applied a predictive model of pedestrian traffic to Melbourne, Australia.
From ACM TechNews