-
limited to) performance tracing for improved scalability, energy efficiency and fault tolerance in ML training / inference. We seek to improve our AI and machine learning work by bringing in tools to assist
-
, depression, and loneliness, and how mental health vulnerabilities increase susceptibility to polarization. Leveraging network science, NLP, behavioral sensing, and causal inference, the project pioneers new
-
, NLP, behavioral sensing, and causal inference, the project pioneers new methods for detecting and mitigating online harms. Its results aim to inform public health, policy, and technology design
-
tracing for improved scalability, energy efficiency and fault tolerance in ML training / inference. We seek to improve our AI and machine learning work by bringing in tools to assist in training and
-
metrics for reducing the environmental impact of AI and cloud computing Empirically evaluate different parts of the AI lifecycle, including development (training), operation (inference) and use Develop
-
), operation (inference) and use Develop methodologies to estimate the long-term impact of AI on carbon emissions and other environmental indicators Build a software framework to quantify the long-term