Category: AI – This Quote

Rodney Brooks argued that intelligent robots should engage the real world instead of relying only on internal representations for today’s AI systems.

Jascha Sohl-Dickstein and colleagues’s words on diffusion models explain a pivotal AI idea and why its evidence, limits, and responsible use still matter today.

Rishi Bommasani and colleagues named foundation models while emphasizing both their broad influence and incompleteness for today’s AI systems.

Emma Strubell and colleagues quantified the financial and environmental costs of increasingly compute-intensive NLP research for today’s AI systems.

Sara Hooker named the hardware lottery to show how infrastructure shapes which AI ideas succeed for today’s AI systems.

Cynthia Dwork and colleagues proposed fairness through awareness by comparing how similar individuals are treated for today’s AI systems.

Latanya Sweeney investigated how online advertising could produce racially associated differences in arrest-related ads for today’s AI systems.

Quoc Le and colleagues explored whether large neural networks could learn high-level features without labels for today’s AI systems.

Jeffrey Dean and colleagues built DistBelief to distribute the training of exceptionally large neural networks for today’s AI systems.

Alex Graves and colleagues applied deep recurrent networks to speech recognition without frame-level alignment for today’s AI systems.

Yoon Kim showed that a simple convolutional network over word vectors could classify sentences effectively for today’s AI systems.

Chelsea Finn and colleagues designed model-agnostic meta-learning so models could adapt from a few examples for today’s AI systems.