The Carbon Footprint of Artificial Intelligence
Looking for ways to cut the release of greenhouse gases attributable to AI use, at a time that usage is very likely to grow.
Artificial Intelligence and Machine Learning
Advertisement
Technical Perspective: Can We Uncover Private Backbone Infrastructures?
In "A Manifold View of Connectivity in the Private Backbone Networks of Hyperscalers," Salamatian et al. assemble techniques from the two broad strategies developed by the networking community and complement them with new ones using concepts from Riemannian geometry.
A Manifold View of Connectivity in the Private Backbone Networks of Hyperscalers
We present a new empirical approach for elucidating connectivity in privately owned and operated backbone networks.
The Principles of Data-Centric AI
Uniting data-centric perspectives and concepts to trace the foundations of DCAI.
GenAI: Giga$$$, TeraWatt-Hours, and GigaTons of CO2
It has become clear that AI's carbon emissions, lifecycle carbon, and other negative environmental impacts are growing explosively.
"Write a 700-word obituary for Vint Cerf." I tried this "prompting" text on a publicly available chatbot. The output? There was a considerable degree of invented but false statements about my life and work.
L-Space and Large Language Models
Design fiction is an approach to understanding and speculating about alternate futures. This column is a piece of such speculative fiction, set in 2025.
Stop Judging AI Using Human Exams
Computer scientists should inform the public that human tests are not a valid way to judge the quality of an AI model, nor a good way to compare AI models to human experts. Computer scientists should stop using human tests to measure AI.
When we ask how deep learning works, are we asking which node represents the answer to "animal, vegetable, or mineral?"
A test should show whether Bard is more powerful than ChattGPT or AI-powered Microsoft Bing.
Automated Evolution Tackles Tough Tasks
Reinforcement learning groups unlabeled data into sets of likes, with the goal of maximizing the cumulative rewards it receives from a human-wrought evaluation function.
Artificial Intelligence-based companions and chatbots allow people to form deep connections through long-term interactions.
AI Bias: Challenges and Solutions
Where does bias in artificial intelligence come from; once we find it, how can we reduce or eliminate it?
VOT Challenge: Computer Vision Competition
VOT Challenge is one of the outstanding competitions in computer vision. At the end of this important event this year, I would like to share my opinions about this competition.
Accelerating Optical Communications with AI
Communications efficiency helps bring photonics to artificial intelligence.
How do we keep track of the truth when bots are becoming increasingly skilled liars?
Looking for truth, when "it's becoming surprisingly easy to distort reality."
Legal Challenges to Generative AI, Part I
Questioning the legality of using in-copyright works for training data and producing outputs derived from copyrighted training data.
Fragility in AIs Using Artificial Neural Networks
Suggesting ways to reduce fragility in AI systems that include artificial neural networks.
Data Science–A Systematic Treatment
While the success of early data-science applications is evident, the full impact of data science has yet to be realized.
Fusing Creativity and Innovation in Asia’s Manufacturing Industry
The Kawahara Universal Information Network Project explores the technologies needed for the next generation of the Internet of Things and HCI.
Achieving Green AI with Energy-Efficient Deep Learning Using Neuromorphic Computing
A description of the effort to develop an end-to-end neuromorphic computing solution in Singapore, including its four aspects of hardware, middleware, software, and system integration.
Human-AI Cooperation to Tackle Misinformation and Polarization
We describe a significant emerging trend away from the techno-solutionist approach to problems that seeks to create and understand a new paradigm: a productive interplay between algorithms and people.
Shape the Future of Computing
ACM encourages its members to take a direct hand in shaping the future of the association. There are more ways than ever to get involved.
Get InvolvedCommunications of the ACM (CACM) is now a fully Open Access publication.
By opening CACM to the world, we hope to increase engagement among the broader computer science community and encourage non-members to discover the rich resources ACM has to offer.
Learn More