Saleh Soltan
@SalehSoltanPrincipal Applied Scientist @Amazon AGI |Ph.D. @Columbia 2017 Views of my own.
Similar User
@ACMSIGCOMM
@ZeyuanAllenZhu
@angcharlesli
@roireichart
@AdapterHub
@umicharmlab
@gil_zussman
@CS3_NSF
@gentaiscool
@ish_iitk
@TingjunChen
@BorisHanin
@shreyansh_26
@Mahshid95521230
@jnsgehring
Amazon has launched Nova, a highly competitive family of foundation models. Nova Pro, Lite and Flash set new standards for the intelligence that can be accessed at the price and speed these models are offered at. Nova Pro, the flagship model, ranks amongst the leading frontier…
It takes a great team to build strong models, but it takes an exceptional team to build state-of-the-art models. I'm incredibly proud of what we've achieved as a team in bringing Nova models to our customers. To many more models to come 🍾 #nova #Amazon aws.amazon.com/ai/generative-…
The ddl is Oct. 1st! Here is the website sites.google.com/view/neurips20…
Our fine-tuning workshop@NeurIPS’24 @neurips24fitml has the following amazing speakers and panelists! Welcome to submit your work and participate in! This workshop aims to explore the frontiers of knowledge of how to efficiently handle nonlinearity!
Our fine-tuning workshop@NeurIPS’24 @neurips24fitml has the following amazing speakers and panelists! Welcome to submit your work and participate in! This workshop aims to explore the frontiers of knowledge of how to efficiently handle nonlinearity!
Happy to announce the #NeurIPS2024 Workshop on Fine-Tuning in Modern Machine Learning: Principles and Scalability! @neurips24fitml For details: sites.google.com/view/neurips20… Organize with @Grigoris_c @BeidiChen @BurkholzRebekka @SalehSoltan @AngelikiGiannou Masashi, @CevherLIONS
Can't agree more
Francois Chollet: OpenAI has set back the progress towards AGI by 5-10 years because frontier research is no longer being published and LLMs are an offramp on the path to AGI
We should call models like Llama 3, Mixtral, etc. “open-weight models”, not “open-source models”. For a model to be open-source, the code and training data need to be public (good examples: GPT-J, OLMo, RedPajama, StarCoder, K2, etc.). Weights are like an exe file, which would be…
Foundation model development is a new focus area for this cycle. Please consider applying. amazon.science/research-award…
Amazing collaboration between @BMW and @alexa99 AI (led by my colleague @samridhishree) to use the LLMs trained by our team to help customers navigate the features in their car easier! youtube.com/watch?v=ZXCjw_…
Now I can freely say what we have been tirelessly working on in the past few months and more. Super excited about the launch!
Amazon’s assistant now has its own Alexa LLM poised to improve the brains in your smart home thanks to its extensive knowledge of smart home APIs. trib.al/5Gce4o3
In case there was any doubt, LLaMA 2 is NOT open source as confirmed by the Open Source Initiative. As much as I appreciate Meta's openness compared to others, I find this misleading marketing a bit irritating.
Congratulations but please watch your language: The license authorizes only some commercial uses. The term Open Source has a clear, well understood meaning that excludes putting any restrictions on commercial use. See `2. Additional Commercial Terms` github.com/facebookresear…
I'm supporting the international protest to stop a proposal for an intensive octopus farm in the Canary Islands. For my video, and much more: pacma.es/actualidad/sto…
Tragic news out of Turkey and the surrounding region. We’re activating Amazon’s logistics and delivery network to quickly get donated relief items such as blankets, tents, heaters, food, diapers, baby food & medicine to those in need. aboutamazon.com/news/community…
We have published a blogpost summarizing some of our recent efforts on using LLMs for generating synthetic annotated data: @AmazonScience amazon.science/blog/using-lar…
Now that AlexaTM 20B is publicly released, you can use it to generate multilingual data for Semantic Parsing (e.g., mTOP). Check our paper for the prompt and results (to appear in AACL 2022): arxiv.org/abs/2210.07074
We just released AlexaTM 20B on Amazon SageMaker Jumpstart. You can now deploy and run inference with the AlexaTM 20B model through JumpStart APIs: aws.amazon.com/blogs/machine-…
When people say GPT-3, do they mean the original GPT-3 or InstructGPT? And which version? It makes a huge difference, so it'd be nice to explicitly specify davinci, text-davinci-002, etc. when making a claim about GPT-3.
🚨 New preprint! 🚨 We refine least-to-most prompting and achieve sota on CFQ (95% accuracy), outperforming previous fully supervised methods. Joint first author work with the formidable Nathanael Schärli.
Check out our new work on generating annotated data for NLU for completely new domains and intents by fine-tuning a large-scale seq2seq model (AlexaTM 5B). Full paper (to appear in @coling2022): arxiv.org/pdf/2209.09900…
Thanks @JeffBezos for sharing our work!
United States Trends
- 1. Remy 11,6 B posts
- 2. $LINGO 35,5 B posts
- 3. Papoose 4.074 posts
- 4. Good Thursday 29,8 B posts
- 5. Zuck 4.441 posts
- 6. #manasol N/A
- 7. Claressa 2.726 posts
- 8. Happy Friday Eve N/A
- 9. #playstationwrapup 1.445 posts
- 10. Our Lady of Guadalupe 5.790 posts
- 11. #ThursdayThoughts 2.026 posts
- 12. #ThursdayMotivation 5.184 posts
- 13. Person of the Year 121 B posts
- 14. Gukesh 40,6 B posts
- 15. #ThursdayMorning N/A
- 16. Wordle 1,272 X N/A
- 17. Dickie V N/A
- 18. Rejected 39,5 B posts
- 19. Voice of America 34,9 B posts
- 20. New York Stock Exchange 26,8 B posts
Who to follow
-
ACM SIGCOMM
@ACMSIGCOMM -
Zeyuan Allen-Zhu | I'm NOT @ NeurIPS 2024
@ZeyuanAllenZhu -
Ang Li
@angcharlesli -
Roi Reichart
@roireichart -
AdapterHub
@AdapterHub -
ARM Lab at University of Michigan
@umicharmlab -
Gil Zussman
@gil_zussman -
Center for Smart Streetscapes
@CS3_NSF -
Genta Winata
@gentaiscool -
Ish Jain
@ish_iitk -
Tingjun Chen
@TingjunChen -
Boris Hanin
@BorisHanin -
Shreyansh Singh
@shreyansh_26 -
Mahshid Ghasemi
@Mahshid95521230 -
Jonas Gehring
@jnsgehring
Something went wrong.
Something went wrong.