MIT - Data to AI Lab

Excellent article in @Forbes today calling #syntheticdata “an all-too-rare example of…genuinely useful” generative AI, for the particular application of software testing. Read @jpwarren profile of @datacebo and @kveeramac : forbes.com/sites/justinwa… #bigdata #syntheticdata #generativeai #data #datascience #enterprisedata #tabulardata #predictiveAI #machinelearning #ML #generativemodels #MLmodels #productiontesting #softwaretesting #cybersecurity #hacking #infosec #security

Young scientists regularly ask me for career advice. Academia or industry? Big company or startup? US or Europe?  Good scientists in AI disciplines are fortunate to have many choices. But choosing can be stressful. I always give the same advice. 1/10

#otd in 2004 Mark Zuckerberg gave his first-ever TV interview, where he said that his company “Thefacebook” had reached 100K users. “Who knows where we’re going to next?” bit.ly/4aOlDM5 (v/@CNBC)

Great article from @thenewstack about open source to business highlighting @datacebo a spin out from DAI lab.

Congratulations @XuLeonard for having his paper on generating #syntheticdata using conditional GAN 1000 citations!

Startup Spotlight: @DataCebo: clouddon.ai/startup-spotli… < checkout my post & video interview with @kveeramac to learn more about how DataCebo uses GenAI to generate high quality data at scale. #SyntheticData #GenAI

A throwback picture from 2018.

CTGAN is a collection of #deeplearning #syntheticdata generators developed at MIT's @lab_dai and one of the most popular models in our #opensource SDV library. Read the what, why & how here 👉ow.ly/E6sy50MOHej #machinelearning

There is a reproducibility crisis in #healthcareAI — but it's not about #datasharing or #privacy. Here are some key learnings we found when developing a structured model-building process with automated tools 👉 ow.ly/jqiT50MHkzG #autoML #machinelearning #automation

We had much to #celebrate in 2022: Alicia Sun (PhD SES) & Lei Xu (PhD EECS) (both toasting in top left image) successfully defended their doctoral theses. AND we were finally able to celebrate Micah Smith (PhD CS)’s 2021 @MIT graduation in person w/ @kveeramac #PhDCelebration

Part of our mission is to identify #ML usability challenges. End-user comprehension is a major one. So we’ve developed a taxonomy to craft features non-#AI experts can understand & become more comfortable using #machinelearningmodel outputs. ow.ly/YRLn50Mw7g2 #ExplainableAI

#Diversity, #equity, #inclusion AND #machinelearning! Way to go @MIT_SCC for encouraging the next generation of #ai pioneers. news.mit.edu/2023/break-thr…

What’s in store for #AI in 2023? The #OpenSource revolution where startups and academia could (and should!) become the epicenters for fundamental AI research, no longer beholden to #BigTech. technologyreview.com/2022/12/23/106…

#SeasonsGreetings from @kveeramac and the team at the #MIT Data-to-AI Lab #AIforGood

Read the @VentureBeat article explaining our recently published research on creating high-quality sentences using #LLMs to train better robust #machinelearning models for language tasks. 👉 venturebeat.com/ai/how-mit-is-… Read our research paper here 👉 aclanthology.org/2022.findings-…

🥳Our very own #MIT Data-to-AI Lab alum Zara Perumal has just been named to the 2023 @Forbes #30under30 for #EnterpriseTechnology. Congrats, Zara! 🎉🍾Looking forward to seeing what great things you'll accomplish in 2023! 💫🚀 ow.ly/u6ji50LRjmp

It’s not always about data science & algorithms...But it’s always about camaraderie & collaboration. Last week, over bowling & drinks, @kveeramac and the @lab_dai team bid farewell to Mihir Thalanki (2nd from right), who was visiting us from @bitspilaniindia. #teamspirit #MIT

We made it! After a year and half of no outing. Thanks @AlexandraZytek for organizing the hike and leading the way. And welcome back to Boston @lab_dai alum @n4atki and Katharine Xiao.

Great chatting with @MattGoodingTM about privacy and how synthetic data can help. Did not get a chance to go into details, but good to highlight that one should avoid broad statements - this is good for privacy, this is not…

We love a good data science competition, but a winning, all-purpose ML model does not exist in the real world. Our users have taught us about trade-offs & we evolved to provide more choices. See our latest #syntheticdata article by @n4atki. bit.ly/feedback-models