Emma Strubell
Assistant Professor, Language Technologies Institute Carnegie Mellon University
- Pittsburgh PA
Emma Strubell's research focuses on efficient and equitable natural language processing (NLP).
Biography
Their work has been recognized with a Madrona AI Impact Award, best paper awards at ACL and EMNLP, and in 2024 they were named one of the most powerful people in AI by Business Insider.
Areas of Expertise
Media Appearances
Emma Strubell | AI Power List
Business Insider online
2024-10-24
In front of a whiteboard from a classroom at Carnegie Mellon University, Strubell explains the Jevons effect, in which the gains from increased efficiency of a technological tool could be negated as use increases. The concept has come into focus as the conversation shifts to efficiency, AI, and the environment. Though they are a proponent for the advancement of AI, Strubell told Business Insider. "GenAI training is a nightmare for energy providers." Their work as an assistant professor at Carnegie Mellon's Language Technologies Institute asks students and researchers to examine the systems that power AI to discover more efficient and environmentally friendly raw materials that power AI.
Greater, newer AI models come with environmental impacts
Marketplace online
2024-06-13
Emma Strubell of Carnegie Mellon University explains why carbon emissions increase with more AI data centers and more powerful AI features.
An AI's Carbon Footprint Is 5 Times Bigger Than a Car's
Popular Mechanics online
2019-06-06
The act of training a neural network, according to the study led by Emma Strubell of the University of Massachusetts Amherst, creates a carbon dioxide footprint of 284 tonnes—five times the lifetime emissions of an average car.
Media
Social
Industry Expertise
Education
University of Maine
B.S.
Computer Science
UMass Amherst
Ph.D.
Event Appearances
AI and the Environment: Sustaining the Common Good
2024 | Markkula Center for Applied Ethics and Next 10 Santa Clara University
Articles
Efficient and equitable natural language processing in the age of deep learning (dagstuhl seminar 22232)
Dagstuhl Reports2023
This report documents the program and the outcomes of Dagstuhl Seminar 22232" Efficient and Equitable Natural Language Processing in the Age of Deep Learning". Since 2012, the field of artificial intelligence (AI) has reported remarkable progress on a broad range of capabilities including object recognition, game playing, speech recognition, and machine translation. Much of this progress has been achieved by increasingly large and computationally intensive deep learning models: training costs for state-of-the-art deep learning models have increased 300,000 times between 2012 and 2018.
Efficient methods for natural language processing: A survey
Transactions of the Association for Computational Linguistics2023
Recent work in natural language processing (NLP) has yielded appealing results from scaling model parameters and training data; however, using only scale to improve performance means that resource consumption also grows. Such resources include data, time, storage, or energy, all of which are naturally limited and unevenly distributed. This motivates research into efficient methods that require fewer resources to achieve similar results. This survey synthesizes and relates current methods and findings in efficient NLP. We aim to provide both guidance for conducting NLP under limited resources, and point towards promising research directions for developing more efficient methods.
Making scalable meta learning practical
Advances in Neural Information Processing Systems2023
Despite its flexibility to learn diverse inductive biases in machine learning programs, meta learning (ie,\learning to learn) has long been recognized to suffer from poor scalability due to its tremendous compute/memory costs, training instability, and a lack of efficient distributed training support. In this work, we focus on making scalable meta learning practical by introducing SAMA, which combines advances in both implicit differentiation algorithms and systems. Specifically, SAMA is designed to flexibly support a broad range of adaptive optimizers in the base level of meta learning programs, while reducing computational burden by avoiding explicit computation of second-order gradient information, and exploiting efficient distributed training techniques implemented for first-order gradients.
Light bulbs have energy ratings—so why can’t AI chatbots?
Nature2024
The rising energy and environmental cost of the artificial-intelligence boom is fuelling concern. Green policy mechanisms that already exist offer a path towards a solution.
A view of the sustainable computing landscape
Patterns2025
This article presents a holistic research agenda to address the significant environmental impact of information and communication technology (ICT), which accounts for 2.1%–3.9% of global greenhouse gas emissions. It proposes several research thrusts to achieve sustainable computing: accurate carbon accounting models, life cycle design strategies for hardware, efficient use of renewable energy, and integrated design and management strategies for next-generation hardware and software systems. If successful, the research would flatten and reverse growth trajectories for computing power and carbon, especially for rapidly growing applications like artificial intelligence. The research takes a holistic approach because strategies that reduce operational carbon may increase embodied carbon, and vice versa.
