Head of ML Strategy, CTO Office, Bloomberg; Former Adjunct Assoc Prof at NYU Center for Data Science (2015-2021); Toronto based; views expressed here are my own
Toronto
Joined October 2007
- Tweets290
- Following681
- Followers1.2K
- Likes739
David Rosenberg retweeted
We are looking for Ph.D. fellows!
Bloomberg will fund research and mentor outstanding graduate students across a wide range of Computer Science and AI topics.
If you want to be considered, please apply by December 14.
Link with information and application details below 👇
David Rosenberg retweeted
We are looking for excellent PhD students across many topics for our Bloomberg CTO AI Research internship next summer.
Link to apply below.
David Rosenberg retweeted
🚀 New paper on evaluating retrieval robustness – how well LLMs handle imperfect retrieval:
1️⃣ RAG >= non-RAG?
2️⃣ More docs >= fewer docs?
3️⃣ Sensitivity to doc order
▶️ 11 LLMs × 3 prompting strategies
Findings: LLMs show surprisingly high robustness—but limitations remain. 1/2
David Rosenberg retweeted
We’re excited to announce @Bloomberg's Visiting Faculty Program, a unique chance to collaborate with our researchers on #AI for #finance & tech. Gain access to world-class data & have real-world impact.
Apply for sabbaticals & part-time roles: bloom.bg/42yDcOh
#research
Thanks @CFAToronto for including me - it was a really nice discussion!
David Rosenberg (@drosen), Head of #ML Strategy in our CTO Office, joins a panel discussion tonight (5 PM EST) hosted by @CFAToronto about "#AI in the Capital Markets," together with Armando Benitez (@jabenitez) of @BMO & Selina Wang of @RBC
bloom.bg/436KjOB
#AIinFinance
David Rosenberg retweeted
In today's 7th BlackboxNLP Workshop at #EMNLP2024, research by @icsaAtEd's Jordi Armengol-Estapé, Lingyu Li, @sebgehr, Achintya Gopal, @drosen, @gideonmann & @mdredze will highlight a method to statically locate topical knowledge in the weight space of #LLMs
#NLProc #AI
David Rosenberg retweeted
During today's #SyntheticData and #GenAI in Finance Workshop at @TheOfficialACM's #ICAIF'24, @drosen, our Head of #ML Strategy, will deliver an Invited Talk about "Private Synthetic Data: Perspectives on Traditional & #DeepLearning Approaches" (10 AM EST)
bloom.bg/4hO5kCU
David Rosenberg retweeted
🚨Summer AI Research Internships🚨
The @TechAtBloomberg office of the CTO is looking for summer AI research interns! ML, NLP, Vision, LLMs
Spread the word.
Interested candidates should apply here:
forms.gle/niB9ZU6eEi6iEznw7
David Rosenberg retweeted
At #TPDP2023, Rodrigo Castellon presented "DP-TBART: A Transformer-based Autoregressive Model for Differentially Private Tabular Data Generation," a paper written w/ our #ML Strategy team's Achintya Gopal, Brian Bloniarz & @drosen
bloom.bg/3PXKdSM
#DifferentialPrivacy
ALT Rodrigo Castellon of Stanford University's Computer Science department co-authored "DP-TBART: A Transformer-based Autoregressive Model for Differentially Private Tabular Data Generation" during his internship with Bloomberg's ML Strategy team in the Office of the CTO.
David Rosenberg retweeted
Our Head of #ML Strategy, @drosen, joined @business' Deputy Editor-in-Chief @RetoGregori in hosting a delegation of Swiss executives visiting #NYC this week for a discussion on how #AI's being used in @TheTerminal today & the potential for #LargeLanguageModels in finance/business
Our new work on synthesizing tabular data with differential privacy. Deep learning methods traditionally underperform marginal-based methods, at least when anybody makes a fair comparison. Our GPT-style models close the gap substantially, and outperform for certain distributions.
Traditional marginal-based methods have long reigned superior for differentially private synthetic tabular data generation. Our #AI researchers show #deeplearning can be competitive & even provide unique advantages!
bloom.bg/3O574cG
#ML #SyntheticData #DifferentialPrivacy
I'm really excited to speak about BloombergGPT at the 7th annual @TMLS_TO! Register at eventbrite.ca/e/toronto-mach…
BloombergGPT is a 50-billion parameter LLM that was purpose-built from scratch. We're proud to welcome @drosen, co-author of BloombergGPT: A Large Language Model for Finance at #TMLS2023 this June 13-14. Come learn more about this fantastic work! @TechAtBloomberg
If you've got a good explanation for this odd behavior of Layer 1 LayerNorm in an early run of BloombergGPT, we may have a #job in #Toronto for you with @TechAtBloomberg's #ML Strategy team (CTO Office).
careers.bloomberg.com/job/de… #NLProc #AIxw
Do you spend a lot of time thinking about attention entropy collapse, AdamW issues, extra layer norms, and pesky loss spikes? If so, we may have a #job in #Toronto for you with @TechAtBloomberg's #ML Strategy team (CTO Office).
careers.bloomberg.com/job/de…
#NLProc #LLMs #AI
Bloomberg’s Ph.D. Fellowship program now accepting applications for 2023-2024!
#CFP: Applications for Bloomberg's 2023-24 #DataScience Ph.D. #Fellowship are now being accepted -- submission deadline: April 28, 2023 (11:59 PM AoE)
bloomberg.com/company/values…
#MachineLearning #ML #NLProc #AI #informationextraction #annotation #humancomputation #speech
David Rosenberg retweeted
🚨NY has a critical need for PPE including gloves, gowns & masks🚨
We need companies to be creative to supply the crucial gear our healthcare workers need. NY will pay a premium and offer funding.
Need Funding? 📞212-803-3100
Have Unused Supplies? 📞646-522-8477
SHARE widely.
David Rosenberg retweeted
Stay at home as much as possible. Listen to the experts, ignore the morons (foreheads). We will get through this together.
David Rosenberg retweeted
Today at FLAIRS-32 (3:15 PM), NLP Architect @astent will present research about stock return predictions co-authored by former interns Ran Zhao & Yuntian Deng, CTO Office colleagues @mdredze & @drosen + our Head of Quant Research Solutions, @ArunVermaQuant bloom.bg/2HtGedd
David Rosenberg retweeted
The @NYUDataScience undergraduate major has been approved by @NYSEDNews! We have already begun to offer undergraduate courses for those looking to minor. You can check out our course selection on the @NyuAlbert "Public Course Search."