Rahul G. Krishnan

4 months of struggle and yet no interview calls? This was the situation of our mentee, who has landed a Data Scientist role! As a CSE graduate, Vinod was struggling to break into the data field. The self-preparation was overwhelming - Jumping between different courses, lacking a structured roadmap, and not knowing which skills to prioritize for interviews. That's when Vinod decided to take help from mentor Tajamul Khan who: ✨ Revamped his resume + profile ✨ Shared a study plan tailored to his goals ✨ Provided constant support, solving doubts 1-on-1 The biggest impact? Vinod landed a job offer within just 1 month of mentorship! Your turn now. Take the smarter route like Vinod did by connecting with a mentor and fulfilling your dream of cracking the Data Scientist role. Start with a FREE 1:1 trial session with your selected mentor. - https://lnkd.in/dEhArNMp Do you still want to take the longer route that is full of struggle and no results? We hope not!🌟

12

1 Comment

Congratulations to Prof Geoffrey Hinton of University of Toronto and John Hopfield for winning the The Nobel Prize in physics for 2024. Soon after the award being announced, the two scientists have warned that AI is a double edged sword. It can benefit the welfare of mankind but it can also cause the collapse of civilization. As they say, the danger emanates from the unknown. One particular danger arises from the interface between AI and the weapons of mass destruction. A small group of countries foresee the danger and want to regulate AI, expand the scope of AI safety, and have multilateral treaties for ethical governance of AI. A second group of countries understand the danger but they want to withhold from a multilateral governance regime in order to bargain a deal on AI in exchange for larger political stakes. The third group of countries do not want to talk about AI safety at all. They want to focus on AI for development and innovation, and suspect that the talk about AI safety is a trap set by the first group of countries to create a monopolistic and discriminatory world order. The third group of countries do not have capacity of producing anything near General AI in the next few decades, indeed not even of manufacturing the chips required for AI. You can speculate which countries belong to which of the three groups. But what we need is a global regime for AI which will take care of the concerns of all the three groups. It is a tough call, but like many tough calls, there is no alternative. Future of Life Institute (FLI) Anthony Aguirre Max Tegmark Rufo Guerreschi Jean-Marc Rickli Simon Cleobury Indrajit Gupta Vivek Y. Kelkar, Oxford Internet Institute, University of Oxford Sir Graham Watson Nicholas Parker https://lnkd.in/dmn5ABQ9

34

3 Comments

We have figured out a lot using mathematics, physics and chemistry to understand nature. We know how to predict electron flow in semiconductors, fluids in tubes, clouds in the sky, airflow over wings, space storms etc. Tools like computational fluid dynamics have remarkable applications in unexpected places. We also have ideas around PINN and we can use those insights to help us understand what happens in a foundational model. And I am sure there are lots more. None of these software artifacts should ever be said to be un-knowable or a "blackbox". That's intellectual laziness. These are all man made structures. We've figured out many parts of nature using math, physics and chemistry - we've mapped the Human Genome, so what's the excuse in not understanding foundational models? I look forward to people recognizing that there are existing human inventions (math, physics etc) whose insights can be used to remove/reduce the probabilistic nature of foundational models and thereby benefiting humanity as a whole. A press conference was held by University of Toronto with Geoffrey Hinton after the announcement that he was one of the two recipients of this year's Nobel Prize for Physics. Geoffrey is probably the only Turing Award winner to also have won a Nobel Prize. And the video - https://lnkd.in/gVHewNCd. Geoffrey is spot on where he says, at about 3:38: "I'm particularly proud of the fact that one of my students fired Sam Altman." Brilliant! #nobelprize #ai #opensource #turingaward

6

3 Comments

One good news: one of my research papers gets accepted in ICSE-2025 (research track). The title of the paper is “Ranking Relevant Tests for Order-Dependent Flaky Tests”. #ICSE2025, #LLM, #testing, #ottawa - #Canada

89

20 Comments

💡 The recently concluded Machine Learning for Healthcare conference (MLHC-2024) in Toronto was incredibly enlightening. ➡ While I've known about it since its inception in 2016, I always opted for IEEE EMBC instead. Now, having experienced both, I believe this conference is a better fit for folks who are specifically interested in #AI or #machinelearning in #healthcare. ➡ Here's why: 1️⃣ Focus: The conference offers a deeper dive into machine learning in healthcare compared to the broader scope of EMBC. 2️⃣ Oral Presentations: All oral presentations are delivered by invited speakers who are group leads. These experts provide insightful overviews of substantial bodies of work from their respective labs, offering a comprehensive view of cutting-edge research in the field. 3️⃣ Research Showcase and Networking: Besides oral presentations, dedicated poster sessions provide an opportunity to hear about current #research from researchers at all levels. The conference's smaller size fosters closer and deeper interactions and engagement. ➡ The conference's smaller size created a more intimate atmosphere. With a single-track format, I had ample opportunity to engage with presenters. ⭐ On a personal note, it was a privilege to finally meet Professor Nigam Shah from Stanford in person. I've admired his work for some time but never had the opportunity to connect before. The review paper on large language models in the context of electronic health records, from his lab (doi: 10.1038/s41746-023-00879-8), is highly recommended! I asked him about the need for a new approach to machine learning in healthcare given the challenges of translating research into real-world impact. I was curious about his thoughts on rethinking PhD programs in this area. You can find the full Q&A in the video below.

40

2 Comments

Today’s Nobel Prize in Physics Congratulations to Geoffrey Hinton for being a co-recipient of the 2024 Nobel Prize in physics for his work in Machine Learning. Some articles this morning have coined him as the grandfather of AI.   Dr. Hinton is a favourite topic of Canadian IP pundits. By all accounts his AI work at UoT was groundbreaking and, along the way, patents associated with some of it are now owned by Google. It is here that the pundits entered the picture with chatter about the loss of Canadian IP.   I too entered the chatter in 2020, sort of. Okay, it was more like talking to myself. At that time I penned a series of three articles about IP in Canada. The first or these (https://lnkd.in/esE7eKc) considered the loss of technology and IP. I said we need to identify the technology that is being developed and patented and we need to invest.  We need to be on the same playing field as those buying Canadian IP. This is one of my motivations for founding IPmart.   A September 15th Opinion in the Globe & Mail framed the IP discussion in terms of productivity. The authors wrote “… we need to dramatically increase the rate at which we protect and commercialize new ideas and the intellectual property (IP) that often starts in publicly funded research.” Hmmm … that sounds familiar! #technology #patents #CDNtech #IPmart

8

“AI was supposed to revolutionize 💊drug discovery. It hasn’t.” #IrrationalExuberance Interesting cover story (above the fold) by Sean Silcoff and Joe Castaldo at The Globe and Mail: https://lnkd.in/gCeVzMcc 💡IMO There needs to be better awareness and expectations (and ethics) around #AI innovation (and more 🇨🇦 Canadian innovation positivity). This is #DeepTech impact research where investor expectations of time horizons need to be in decades (not yesterday) and where exploration is seen as impactful learning (positive) rather than another Canadian startup failure (negative). “One of the promises of machine learning was better drugs, faster. A decade in, AI has failed to live up to the hype.” The general public has no idea the computing evolution that started as machine learning and has now evolved into computers capable of artificial intelligence (AI), generative AI, and quantum computing. #Quantum computers (basically all knowing god-like computing seen in the show Westworld) will achieve what would take AI computers decades of scientific AI research development, and do it in seconds. So does failed expectations mean that current research is a failure? No. We simply need more accurate expectations. Research and innovation requires time, a ton of money, and - most importantly - optimism. Hope. 💡🔥🚀 #Innovation #CorporateInnovation

9

1 Comment

Computer science professor at University of California, Berkeley : Stop pursuing CS = You are welcome to study CS for Complexity slaying of Computational challenges but not for jobs. Here is what Prof. James O'Brien said: I've been a professor of computer science at UC Berkeley since 2000. Recently, I've noticed how difficult it is for tech graduates to enter the job market right now. At least in technology, it's a very different landscape than it has been in the past. It's not just my students; it's a hard time for everyone right now. Over the years, students I've worked with, taught, or known from other schools have typically had little trouble finding a job after graduating with a tech degree. There were lots of opportunities. Quite often, recent graduates would receive multiple offers and be left to try to decide between two really great choices. But now, things have changed. There's still demand for people, and some people are getting jobs, but not across the entire industry. Our view from Berkeley is also a little distorted because we're one of the top schools, so most of our students are still getting offers. But rather than getting five great offers, they might only get one great offer or an offer from their second choice. This is certainly not what we've seen in the past. There are also people coming from other schools that don't have quite the strong reputation that Berkeley does, and they're not getting job offers at all. Some programming roles are disappearing. It's not just the technology jobs that are being affected. Any "hands-on keyboard job" is at risk of being contracted out offshore or outsourced to AI. Another problem is that many companies are also trying to consolidate a bit. These things are all coming together like a perfect storm. In tech, programming positions that aren't that demanding are disappearing. They are still entry-level jobs that involve basically babysitting the AI — but these systems are getting better very quickly. = This matches No Work Theory. What? No one should ever work. WTF?? Work is the source of nearly all the misery in the world. Almost any evil you’d care to name comes from working or from living in a world designed for work. In order to stop suffering, we have to stop working. That doesn’t mean we have to stop doing things. It does mean creating a new way of life based on play; in other words, a ludic revolution. By “play” I mean also festivity, creativity, conviviality, commensality, and maybe even art. There is more to play than child’s play, as worthy as that is. I call for a collective adventure in generalized joy and freely interdependent exuberance. Play isn’t passive. Doubtless we all need a lot more time for sheer sloth and slack than we ever enjoy now, regardless of income or occupation, but once recovered from employment-induced exhaustion nearly all of us want to act. Pursue 007 life based on Universal Basic Income instead of 996 corporate slavery. That is Utopia.

45

37 Comments

The next stop on my journey to attempt to grasp how the recent Nobel Physics prize in AI is considered physics led me to the video of the University of Toronto press conference which featured Hinton. This snippet from Hinton stood out, for me: "I'm not a physicist. I have very high respect for physics. I dropped out of physics after my first year at University because I couldn't do the complicated math so getting an award in physics was very surprising to me. I'm very pleased that the Nobel committee recognized that there's been huge progress in the area of artificial neural nets and Hopfield's work was closely related to physics and some of the early work I did with Terry Sejnowski on Boltzmann machines was inspired by statistical physics but more recently the work has had less relationship to physics and so I was very surprised that I got a prize in physics." There are three key phrases there: 1. "Hopfield's work was closely related to physics". He didn't elaborate, unfortunately. 2. "some of the early work I did with Terry Sejnowski on Boltzmann machines was inspired by statistical physics". So, physics-inspired, but is it physics per se? 3. "but more recently the work has had less relationship to physics". And finally, Hinton says "and so I was very surprised that I got a prize in physics." So, I still have not discovered any compelling justification for a physics prize for AI. #Physics #NobelPrize #AI #ArtificialIntelligence #NeuralNetworks #GenerativeAI #GenAI #LargeLanguageModels #LLM #QuantumComputing #QuantumApplications #QuantumAlgorithms #QuantumInformationScience #QIS #QuantumTechnologies #QuantumTech #Quantum https://lnkd.in/eiD5xA4D

9

🔬 Researchers from Lawrence Berkeley National Laboratory (Berkeley Lab), UC Berkeley, the Institute for Systems Biology, and Université Laval have unlocked new insights into the TIP60 protein complex, a critical gateway to DNA accessibility. Published in the journal Science, this groundbreaking study reveals the atomic structure of TIP60, paving the way for potential breakthroughs in targeted therapies for diseases such as cancer and Alzheimer's. Using advanced high-resolution cryo-electron microscopy, the team elucidated how this complex enables essential DNA repair and transcription processes, which are vital for cellular health and often disrupted in disease states. This research not only enhances our understanding of genetic regulation but also opens exciting avenues for developing therapies that manipulate TIP60's activity to treat genetic and neurodegenerative disorders. Explore how this significant discovery could transform future medical treatments! #Genetics #HealthcareInnovation #ScienceNews Read more: https://lnkd.in/eMnjr2bp

3

Excited to share our DistiLLM accepted at #ICML2024. DistiLLM is a better (performance), faster (training efficiency), and stronger (theorem) distillation method to significantly enhance the performance of SLMs. #efficientAI #slm

26

EMNLP 2024 results are out, we got in three papers, congrats to students :) First paper of many, Vatsal, Pranshu, Suyash, Janit, Kunal, Srija and Adnan, Congrats 🎉 🎊 🍾 🎈 Evaluating Concurrent Robustness of Language Models Across Diverse Challenge Sets (main) - Vatsal Gupta, PRANSHU PANDYA, Tushar Kataria, Vivek Gupta, Dan Roth tldr: make models robust on multiple adversarial challenge set simultaneously  arxiv: https://lnkd.in/ei-65Y6g Knowledge-Aware Reasoning over Multimodal Semi-structured Tables (finding) Suyash Vardhan Mathur, Jainit Bafna, Kunal Kartik, Harshita Khandelwal, Manish Shrivastava, Vivek Gupta, Mohit Bansal, Dan Roth tldr: introduce the novel task of reasoning on multimodal tabular data, text + images in a structure (first of a kind) arxiv: https://lnkd.in/e5ycusfn   Unraveling the Truth: Do LLMs really Understand Charts? A Deep Dive into Consistency and Robustness (finding) Srija Mukhopadhyay, Adnan Qidwai, Aparna Garimella, Pritika Ramu, Vivek Gupta, Dan Roth           tldr: VLM although seem very accurate on chart, are not all robust and consistent in performance arxiv: https://lnkd.in/eFFA-sVX We also got an workshop paper, at BlockBoxNLP 2024 @EMNLP, here the details Enhancing Question Answering on Charts Through Effective Pre-training Tasks Ashim Gupta, Vivek Gupta, Shuo Zhang, Yujie He, Ning Zhang , Shalin Shah, Ph.D. tldr: models is not robust & consistent on charts, can we use pre-training on simple auxiliary tasks to enhance model robustness arxiv: https://lnkd.in/egbRHm-4

386

12 Comments

I successfully attained 2 years of PhD funding! 👩‍🎓 The funding is supported by Mitacs through the #MitacsAccelerate Program. The program supports #PhDEntrepreneurs and #CanadianInnovation - Globally. It took about 2 years, to formalize my research goals. I am grateful for support that I received from my supervisor Dave Mason at Toronto Metropolitan University and my partner Shawn Berney. Grateful for Innovation Boost Zone, the university incubator, where it all started by me participating in the Lab2Market of summer 2020. Key people at IBZ that support me are Wesley Kosiba and Ashu Syal and Torrey Sirdevan! Thank you Guys! I had a great ally Leslie Bone who supported me through the years, especially, through the submission process! Last but not least! Thank you to Steven Delaney, PhD who encouraged me to step-up! With his encouragement I provided a IEEE presentation for the Toronto Chapter on #Sustainable #SupplyChain for #OpenEducation” – when at the time my PhD was focused on #ArtificialIntelligence! That nudge and vote of confidence was instrumental for me when I had to make a tough decision to change my PhD thesis from #AI, which I will always love! into something completely different where I had zero traction. It was a Leap of Faith for me! But because I took the leap of faith I am uniquely positioned in the world of technology to create new economic and social opportunities. As a CTO and cofounder of OpenSailAssets, I am eager to leverage technology for positive social impact. The purpose of my PhD is to figure out how to best use technology to support the #DigitalEconomy and #DigitalSociety and through it unleash the potential that #ArtificialIntelligence can have on #CustomLearningExperiences. My motivation is rooted in personal experiences as a #neurodivergent learner. I would not have been able to believe you if you told the 7-year-old me that my #dyslexia is actually a gift. Though early on, I knew I had great ability to execute complex visual tasks. As such I always got A+ in Art! Sculpting realistic 3-D objects was easy for me. Learning how to read was not. But I always loved to learn and I am endlessly curious! So much has changed since I attended #PrimarySchool. We have far more flexibility to support and fuel independent learning in the #classroom and at home. However, we can do more. For our learners to thrive we need to support their network of #Teachers and their #Parents around #ContentFlexibility that they use to learn from. #CustomLearningExperiences for #LearningAccessibility, #LearningEngagement and #LearningIndepedence. This is the goal of OpenSailAssets. To bring content flexibility to our #YoungLearners as well as #AdultLearners, who stand to benefit from the same ecosystem. If you are #teacher and are interested in increasing content use flexibility, let me know by replying to this post! If you are #researcher, learn about Mitacs funding opportunities: https://lnkd.in/g9J_tugG

27

5 Comments

Excited to connect with fellow researchers at #NeurIPS2024 in Vancouver! We're presenting two posters at the SOLAR workshop today focused on advancing privacy in LLMs. 🔒 First up: A fresh look at the DP-utility tradeoff. It's well known that DP impacts utility - but what does "bad utility" really mean for language generation? We're presenting a taxonomy of errors, showing how we not only recover from privacy-induced issues, but how to surpass non-private models in reducing hallucinations! Curious? Check out our findings here: https://lnkd.in/eTsx3cEs Plus, sharing insights from my IBM Research internship on balancing personalization and privacy in LLM conversations. We take a step towards solving this problem by operationalizing contextual integrity for data minimization in user->LLM information flows. We introduce a framework that empowers users to prevent oversharing.  Explore our preliminary results: https://lnkd.in/ej_HDcTw 🔐 Drop by to discuss these findings and explore the future of privacy-preserving AI! Would love to hear your thoughts.

57

1 Comment

A new collaborator article is out on biostatistics.ca! Thanks to Renee (Hui Xin) Ng for sharing this insightful piece about using sophisticated Machine Learning (ML) algorithms to develop biological constructs related to brain aging. Specifically, the article explains why and how these models are biased, how to address these biases, and what they mean for research in this field. It's an fascinating read, check it out 👇 https://lnkd.in/ehKpzDsu

12

6 Comments

Excited to share our latest publication in Computers in Biology and Medicine. In collaboration with the BC Children's Hospital Research Institute, this study introduces BRI-Net, an interpretable deep-learning model that leverages adversarial training to learn an unbiased representation of data, helping identify genetic markers associated with specific phenotypes. Using genomic data from a cohort of 362 pediatric cancer patients treated with cisplatin, recruited by the Canadian Pharmacogenomics Network for Drug Safety, our model identified two statistically significant SNPs associated with cisplatin-induced ototoxicity: CERS6 rs13022792 and TLR4 rs10759932. To explore our methodology and findings in detail, check out the full paper at https://lnkd.in/gEn4fZb9. A huge thanks to my amazing co-authors and collaborators, especially Bahareh Kashani, for making this possible!

66

4 Comments

Those who talked with me in the past couple years might remember that I have a long-lasting grudge with quantum circuits (https://lnkd.in/gp_snVEY). As it turns out, the problem runs much deeper than just circuits, some other people also noticed it and were happy to have multiple discussions on the topic. The first step was to organize a workshop at last year IEEE Quantum Week (you can read summary here: https://lnkd.in/gPnGz23g) and the next one was writing a paper on the topic. So I'm glad to say that together with Olivia Di Matteo Santiago Núñez-Corrales, Ph.D. Steve Reinhardt and Tim Mattson (and with non-trivial contributions from Tom Lubinski and Yuval Sanders), we've recently posted on arXiv a paper titled: "An Abstraction Hierarchy Toward Productive Quantum Programming". If you're more on the NISQy side of things and feel like writing and running programs on contemporary QPUs is arduous... Or if you're more on the FTQCy side of things and you feel confused about how we will be running programs on future computers... I hope this paper will provide some good food for thoughts! We were not aiming at finding solutions, but rather articulating some of the issues and providing a better framework to thing about these problems. Let us know if we succeeded! https://lnkd.in/ggAHtrPc

152

9 Comments

Go check out Brevitas at the links below and start experimenting with MX and all the latest quantization algorithms!

11