article thumbnail

Visual Blocks for ML: Accelerating machine learning prototyping with interactive tools

Google Research AI blog

Researchers optimize the model with quantitative metrics on a fixed set of data, but real-world performance requires human reviewers to evaluate in the application context. Participants could not quickly and interactively alter the input data or tune the model. Acknowledgements This work is a collaboration across multiple teams at Google.

article thumbnail

Visual captions: Using large language models to augment video conferences with dynamic visuals

Google Research AI blog

We measured the performance of the fine-tuned model with the token accuracy metric, i.e., the percentage of tokens in a batch that were correctly predicted by the model. We used 1276 (80%) examples from the VC1.5K dataset for fine-tuning the large language model and the remaining 319 (20%) examples as test data.

professionals

Sign Up for our Newsletter

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.

article thumbnail

NpTech Summary: Advocacy 2.0, Sketchcastes, and NpTech in Different Languages

Beth's Blog: How Nonprofits Can Use Social Media

As both a lesson and as a metric, failure is potentially productive at every level of socialmarkets - from the repeated return of one homeless person to shelter, to the repeated attempts to attach a value like SROI to such a story. The "Failure is an option" particularly caught my eye.

Nptech 50
article thumbnail

Google at EMNLP 2022

Google Research AI blog

Zhao , Yi Luan , Keith B.

Google 52
article thumbnail

Stanford AI Lab Papers and Talks at NeurIPS 2021

Stanford AI Lab Blog

Powers, Yianni Laloudakis, Sidhika Balachandar, Bowen Jing, Brandon Anderson, Stephan Eismann, Risi Kondor, Russ B. Townshend, Martin Vögele, Patricia Suriana, Alexander Derry, Alexander S. Altman, Ron O.

Contact 40
article thumbnail

How to Improve User Experience (and Behavior): Three Papers from Stanford's Alexa Prize Team

Stanford AI Lab Blog

While there is a large body of prior work attempting to address this issue, most prior approaches use qualitative metrics based on surveys conducted in lab settings. We believe that these metrics measure the effectiveness of a response strategy more directly than user ratings as done in Cohn et al.

Alexa 40
article thumbnail

Google at NeurIPS 2022

Google Research AI blog

Chi The Nature of Temporal Difference Errors in Multi-step Distributional Reinforcement Learning Yunhao Tang, Mark Rowland, Rémi Munos, Bernardo Ávila Pires, Will Dabney, Marc G.

Google 52