AI Researcher - Google, RAIL Co-Founder, Microsoft, WIRED Fellow, Cambridge Uni., MIT Media Lab, Affectiva.
Seattle, WA
- Tweets 353
- Following 395
- Followers 1,372
- Likes 522
Attending @ICCVConference? The CVPM workshop has a stellar line up of speakers from #Apple, #MERL, #Google, #UIUC and also a presentation of the largest and most diverse video blood pressure dataset.
Time⏰: Sunday 8am
Location📷: Rm 328
1
399
Great to be at @CVPR. Looking forward to the 7th (Seventh!) workshop on Computer Vision for Physiological Measurement this afternoon. es.ele.tue.nl/cvpm24/ Talks from @senguptroni and Dr. Jake Sunshine on inverse graphics for health and devices to detect cardiac emergencies.
4
9
2,922
The @ResponsibleAIL License Generator (licenses.ai/rail-license-gen…) is now live. If you have an #AI/#ML project consider using it to generate a license with customized use clauses to prevent harms.
1
4
7
2,909
See our ICML'24 position paper for more information: arxiv.org/abs/2402.05979
1
1
2
616
Particular credit for the license generator goes to Tim Korjakow, Jesse Benjamin, Scott Cambo, @danish_c.
1
399
Language models can increase access to, and the utility of, insights from health data. This work has been a big effort from a multidisciplinary team towards that end. Really please to be able to share our findings.
Today on the blog, read about the latest from our two new research papers on how AI, particularly fine-tuned Gemini models, can create personalized health experiences that cater to individuals’ unique health journeys. →goo.gle/3RnwHbl #AI #healthcare #personalizedhealth
ALT To evaluate Personal Health LLM (introduced in our first paper), we curate three benchmark datasets that span long-form coaching recommendation tasks, assessments of expert domain knowledge, and prediction of self-reported sleep outcomes.
1
19
834
There is a amazing potential for foundation models in robotics but data is a huge bottleneck. Combining simulation and LLMs to create models is a really exciting direction. Can't wait to see the impact of this.
Introducing GRID: the General Robot Intelligence Development platform, designed for prototyping smart and safe robots rapidly using foundation models, LLMs, and simulation.
Paper: arxiv.org/abs/2310.00887
Try now: portal.scaledfoundations.ai
GitHub: github.com/scaledfoundations…
🧵👇(1/N)
1
13
1,518
Daniel McDuff retweeted
Introducing GRID: the General Robot Intelligence Development platform, designed for prototyping smart and safe robots rapidly using foundation models, LLMs, and simulation.
Paper: arxiv.org/abs/2310.00887
Try now: portal.scaledfoundations.ai
GitHub: github.com/scaledfoundations…
🧵👇(1/N)
15
175
823
277,338
This has been a really exciting week for #RAIL (licenses.ai/) - First #Llama2 was released with a RAIL-like license with behavioral usage restrictions and now the @ai2_allennlp (AI2) sets out their position on adopting OpenRAIL-DMs (lnkd.in/grwJ6r_3).
1
3
598
Daniel McDuff retweeted
🌟Announcing the 1st PERception, Decision-making, and REAsoning through Multimodal Foundational Modeling (PerDream) Workshop @ICCVConference ! Join us in fostering extensive research and vibrant discussions on this emerging topic. 🧵👇
1
14
55
14,916
Daniel McDuff retweeted
Today we published a paper that gives an overview of the evolution of licensing choices for #ML models published on @huggingface. We found that over the last 5 months, the use of Responsible AI Licenses (RAIL) has grown from 0.5% to 7.1% of repositories with licensing metadata.
1
7
10
2,154
At @ResponsibleAIL we have been working with @huggingface, @BigscienceW & @RealAAAI to give practitioners more options for distributing code, models, data and apps. If you are submitting to @RealAAAI consider an AI Pubs license. @danish_c, @juliakhaines, @bhecht, @Carlos_MFerr
Responsible AI Pubs Licenses are built on a model for behavioral-use that aims to reduce the risk of negative outcomes and misuse of AI. These licenses are similar to licenses used by BigScience and over 6000+ related repositories for AI models and code. Community feedback is essential, and we are grateful to receive input. Please complete the survey on this page if you plan to use AI Pubs Licenses ➡️ licenses.ai/ai-licenses
1
7
1,126
Daniel McDuff retweeted
Responsible AI Pubs Licenses are built on a model for behavioral-use that aims to reduce the risk of negative outcomes and misuse of AI. These licenses are similar to licenses used by BigScience and over 6000+ related repositories for AI models and code. Community feedback is essential, and we are grateful to receive input. Please complete the survey on this page if you plan to use AI Pubs Licenses ➡️ licenses.ai/ai-licenses
7
24
15,625
One month until the submission deadline for the CVPR'23 Workshop on Computer Vision for Physiological Measurement. If you work in the field of CV/ML and physiological measurement these are one of the best events to contribute to and attend: es.ele.tue.nl/cvpm23/
1
488
Looking forward to giving the Grundfest Lecture @Caltech tomorrow on Computational Imaging and applications in health sensing visual.ee.ucla.edu/web_serie…
1
2
20
4,027
Daniel McDuff retweeted
🚨Announcing CVPR 2023 workshop on Computer Vision for Physiological Measurement🚨
We hope to bring together the CV and health sensing communities, discuss the opportunities, challenges & latest advances for human health sensing with CV/ML.
es.ele.tue.nl/cvpm23
@CVPR #CVPR
2
8
29
13,900
Daniel McDuff retweeted
Excited to share the call for participation from the @ResponsibleAIL initiative. There's work to be done towards developing best practices, tooling, governance, data licensing and more! licenses.ai/blog/2023/1/17/r… #ResponsibleAILicenses #RAIL #AI #ResponsibleAI
This has helped us identify several important directions to move this initiative forward, from tooling to legal work to public advocacy. We would love to have your input and work on these ideas together! Detail: licenses.ai/blog/2023/1/17/r…
1
5
9
2,167
Continuous temporal prediction in CV still seems relatively understudied in my opinion compared discrete classification tasks, so excited to share this work led by @yang_yuzhe on periodic learning with applications in remote sensing and health.
📢Check out our latest work on self-supervised learning of *periodic* information from data!
We present SimPer, a simple SSL regime for learning periodic targets. w/ @xliucs, Jiang, Silviu, Dina, Ming, @danmcduff. Thanks @_akhaliq for sharing!
2
11