Raj Abhijit Dandekar

How will you respond ? if I tell you that Physics is simpler than Biology? Yes, it's easier. Physics is based on a magical learning process. Let me explain it to you using three pointers: ~ Structured Learning Process: Physics has an organized learning process that moves from one approach to the next, which makes it appear more effortless than different topics in biology. ~ Unified Learning: The subject is learned in a unified manner, where fundamental principles apply across different topics, providing a complete understanding. ~ Mastering Basics: Mastering basics is the most important part of understanding physics, such as understanding cross-products will make a solid foundation for understanding complex concepts like torque, angular momentum, conservation of angular momentum, etc. In conclusion, while both physics and biology have their complexities, one can find physics to be simpler due to its structured learning process, unified way of learning, and focus on mastering basics. By embracing this magical learning journey, physics will become more accessible and easier to comprehend. Here is my YouTube link regarding the same :- https://lnkd.in/gAsvwTh6 #NEETPhysics #NEETExam #PhysicsStudents #PhysicsEducation

25

Jianbo Shi and Jitendra Malik's paper on "Normalized Cuts and Image Segmentation" (https://lnkd.in/dcyU5Skr) published in 2000, is a seminal work that introduces an interesting method for image segmentation by using concepts from graph theory and linear algebra. It is one of the most highly cited papers in Computer Vision with over 20k+ citations. By modeling the image as a graph, where pixels represent nodes and the edges between them encode the similarity, the segmentation problem transforms into one of graph partitioning. The segments then can be thought of as subgraphs. The key idea is to partition the graph to minimize the normalized cut (a criterion that evaluates the cut cost: the total weight of the edges removed to split the graph) relative to the total edge connections within each group. This problem is then elegantly converted into a linear algebra problem: finding the partition corresponds to solving for the second smallest eigenvector of the Laplacian matrix of the graph. Here's a short report on the entire paper, prepared as part of our project component for the Linear Algebra and its Applications course last semester. Please feel free to comment or address any issues if there are any! #ImageSegmentation #MachineLearning #ComputerVision #DataScience #ScientificResearch #LinearAlgebra #GraphTheory #Report #project

125

5 Comments

I always thought IITs were about quality over quantity, but the BS in Data Science and Applications program at IIT Madras has shattered that myth! This single program is challenging all private universities in enrollment numbers. While private institutes struggle to scale their online offerings, IIT Madras has proven they can do both quality and scale. It's a wake-up call for private universities: it's not just about big numbers in regular modes, but excelling in online education too.#onlineeducation #highereducation #datascience #iitm

9

Thoughts on this? >> Meet ONI: A Distributed Architecture for Simultaneous Reinforcement Learning Policy and Intrinsic Reward Learning with LLM Feedback >> Comment below! >>> lqventures.com #mhealth #healthtech #AI #IoT #industry40

1

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

🚨 Big News for IIT Aspirants: JEE Advanced 2025 Now Allows 3 Attempts! – Here’s What It Means for Your Future 🎓 The recent update by JEE Advanced to allow students three attempts is a game-changer in the IIT landscape. Students from the 2023, 2024, and 2025 batches can now try thrice, providing more chances to secure a spot in India’s top institutes. Here’s how it impacts preparation and career paths. 👇 1. Enhanced Flexibility 🕰️ With three attempts, students can strategize, learn, and improve over time, easing the pressure to succeed immediately. This flexibility helps students feel more in control of their journey. 2. Better Mental Health & Reduced Burnout 🧠 JEE prep is intense. Three attempts provide a safety net, reducing “one-shot” pressure and fostering a more balanced approach, likely leading to better mental health. 3. Encouraging Long-Term Skill Development 📚 This structure supports early, consistent preparation, building skills like time management and resilience that will benefit students beyond IIT. 4. Broadening Access & Diversity 🌏 This policy can attract diverse students. Those from varied backgrounds now have multiple chances to improve, enriching IITs with a mix of perspectives. 5. Balanced Career Planning 🎯 While more attempts can reduce pressure, students must avoid getting “stuck” in repeated cycles. Clear timelines and exploring alternate paths like data science and AI can ensure well-rounded career planning. 6. Impact on the Coaching Industry 📈 The coaching sector may see demand for extended support, leading to a focus on gradual, sustainable learning rather than cramming. Conclusion Allowing three attempts brings both opportunities and responsibilities. Success in JEE is just one step; building a meaningful career requires resilience and openness to diverse paths. 🌟 Make the most of this chance! Focus on learning, stay resilient, and remember, your future holds many possibilities. 💪 #IITJEE #CareerCounseling #EducationPolicy #MentalHealthMatters #FutureReady

5

#2much The increasing pressure on students preparing for competitive exams like IIT-JEE is alarming, but a recent incident highlights just how invasive this pressure can become. A girl has claimed that her parents are using a camera to monitor her study sessions, raising serious concerns about the violation of a teenager’s privacy. Such extreme surveillance invades a young person’s personal space and undermines their sense of individuality. For teenagers, privacy is crucial for their psychological development, and constant monitoring can lead to heightened stress, anxiety, and a feeling of being controlled, ultimately harming their mental well-being and autonomy. https://lnkd.in/gdrT6aTt

1

🚀 Creating IIT-JEE Question Paper using AI Agents The recent release of JEE Advanced results took me back to the day when I nervously checked my own scores. As a fun experiment, I wondered: could I use Generative AI to create an entire JEE question paper from scratch? So, I created AI agents to generate high-quality JEE questions. Here's how I built it: Simply prompting LLMs to give good questions didn't work well because the quality and difficulty level of the questions were inconsistent. I used the CrewAI agent framework and GPT-4o model for this exercise. The process involves two key agents: 1️⃣ The Question Creator Agent crafts thought-provoking questions based on specific topics from the IIT-JEE syllabus. 2️⃣ The Question Reviewer Agent meticulously reviews each question, providing feedback and suggestions to refine them. These agents converse among themselves to generate high-quality, challenging questions that test the analytical skills and conceptual understanding of JEE aspirants. See the video below for a tutorial. I created a detailed Python notebook to show you how you can do it too! (Link in comments) —— Want to learn how to build AI apps? Join me for a free, 1-hour “Function Calling with LLMs” workshop on Friday, where we will explore how to add superpowers to LLMs. Register here: https://lnkd.in/gJwkcJXg #AIAgents #GPT4o #BuildFastwithAI #FunctionCalling #FreeWorkshop 

312

11 Comments

Project specific call for internship: We are looking for 2 summer interns in India in summer 2025 for the Quantum Algorithm and Engineering (QAE) team. The goal of these internships will be to understand domain specific best practices (e.g. error mitigation, qiskit-addons, transpilation) to execute utility-scale projects on IBM Quantum hardware using Qiskit, and bring this knowledge to the community in the form of one or more tutorial notebooks. We are specifically looking for students who have a strong understanding of either (i) quantum for bioscience, or (ii) parallel and serverless computing for quantum middleware. Good grasp of Qiskit 1.0 or higher is necessary. Also note that an intern must be a student and should not graduate in 2025. Please apply for IBM Quantum summer internship in India here: https://lnkd.in/gPfaaxFg If you are interested in any of these two projects, please send me a message over linkedin after applying (or if you applied previously via this link). Also note that (i) sending me a message does not guarantee an internship position, but it will help me during screening, and (ii) I am not aware of other quantum projects for summer 2025, so please don't ask for that. Nate Earnest-Noble #ibm #ibmquantum #quantumcomputing #qiskit

207

26 Comments

🎉 Exciting News from IIT Madras! 🎓 The B.S. in Electronic Systems at IIT Madras is making waves! Featured in The Hindu, this industry-ready course is breaking the mold by not requiring JEE for admission. 🚀 With 1,500+ students enrolled in its first year, here’s why this program stands out: ✅ Flexible entry & exit options ✅ Industry-focused curriculum ✅ Open to anyone with Class 12 Maths & Physics Ready to shape the future? Check out the buzz in The Hindu: https://lnkd.in/gcqqVUZ8 #IITMadras #BSElectronicSystems #Innovation #EdTech #SkillDevelopment #HigherEducation #TheHindu #FutureReady

266

2 Comments

The next guest speaker in my LLM course is Ram Sriharsha, the CTO of Pinecone, who will talk about Retrieval Augmented Generation (RAG) on Monday October 28. ➡️ Register here: https://lu.ma/r7yokyvc Title: Knowledgeable Models Combine Retrieval with Generation: An Introduction to RAG Date: Monday October 28 from 1:45 PM to 3:15 PM ET Abstract: Large Language Models have emerged into mainstream over the last few years. While they have demonstrated some huge successes already, they have also had challenges with hallucinations, difficulty with dealing with new knowledge and in retrieving relevant, accurate knowledge for downstream tasks. In this talk, we will provide a brief overview of the problem, and discuss a promising approach to solving these problems through what is known as Retrieval Augmented Generation (RAG). Alongside, we will also introduce vector databases, what makes them different from traditional databases and how they fit into RAG. Bio: Dr. Ram Sriharsha is the CTO at Pinecone. At Pinecone, Ram is responsible for overseeing the company’s technical direction in both engineering and research. Prior to joining Pinecone, Ram held numerous senior positions at some of the most well-known companies in the tech industry. Most recently, he was Vice President of Engineering at Splunk. Before that, he was a product and engineering lead at Databricks for their unified analytics platform for genomics. Some years after receiving a PhD in Physics from the University of Maryland, Ram was a Principal Software Engineer and later a Principal Research Scientist at Yahoo. Supplementary Readings: ➡️ The original paper on RAG: "Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks" (https://lnkd.in/e_e56Y54) ➡️ Long context considerations: "Lost in the Middle: How Language Models Use Long Contexts" (https://lnkd.in/ehED53Fe) ➡️ Article on contextual RAG from Anthropic AI (https://lnkd.in/ehK4GwMc)

69

🌟 When Should You Start Preparing for IIT JEE? 🌟 Is starting early in Classes 8, 9, or 10 the key to cracking IIT JEE Main & Advanced, or is focused preparation in Classes 11 & 12 enough to succeed? What should be the approach of students currently aiming for class 10 Boards only ? Watch my full video to discover the best approach, expert tips, and how to optimize your preparation journey for this prestigious exam! 🚀 #IITJEE #PreparationTips #EducationMatters #BrothersAcademy

12

Students who are currently having a hard doing research or are feeling stuck during their graduate studies, I have a story to share. Here’s my publication history in undergraduate and PhD: Undergraduate at IIT Madras: 2013 (1st year): 0 publications 2014 (2nd year): 0 publications 2015 (3rd year): 0 publications 2016 (4th year): 1 publication in Journal of Fluid Mechanics. 12 citations until today. That publication in my final year got me admitted into MIT for my PhD. PhD at MIT: 2017 (1st year): 0 publications 2018 (2nd year): 0 publications 2019 (3rd year): 0 publications 2020-2021 (4th, 5th year): 4 publications. 250+ citations until today Those publications in my final year of PhD got me ML job offers, I defended my PhD thesis, got featured in MIT News and collaborated with governmental agencies. There is a clear pattern to this. Research is like the growth of the Chinese Bamboo tree. You take a little seed, water it and fertilise it for years. Nothing seems to happen. During the 5th year, the bamboo tree grows 90 feet in 6 weeks! Research takes a very long time to show results. The first 3 years of both my undergraduate and PhD had no results. If you are patient, the results will come and the results will be very, very impactful. Having respectable publications led to the biggest career transitions for me. It has been the highest return on investment of my career. The investment of countless hours spent in the lab with no results, lab transitions, reviewer rejections and years of working on a problem in isolation. Research is underrated. The impact it can have on your career is not just underrated, it’s neglected. It's hard and it takes time. Hang in there. Results take time to come. When they come, they will be very rewarding. P.S: Here is a picture of me defending my PhD thesis at MIT! One of the happiest days of my life. If you are having tough days with research, I hope this post motivates you to keep going.

517

19 Comments

Vizuara's AI lecture series "The AI Hour" starts tonight. Our first speaker is Satwik Sinha who will deliver a talk on Physics Informed Neural Networks (PINNS). Satwik is a distinguished AI researcher and software engineer with more than a decade of hands-on experience. He is currently pursuing a master's degree in computer science at Illinois Institute of Technology Chicago. He has also delivered guest lectures at University of Southern California (USC). Physics-Informed Neural Networks (PINNs) are transforming how we approach complex scientific and engineering challenges. Regular and partial differential equations (ODEs and PDEs) can be used to solve real world problems. PINNs are a powerful way to do this because they seamlessly incorporate physical laws into neural network training. So why attend this lecture? You will... 1) Learn how PINNs represent solutions to PDEs directly within neural networks. 2) Discover how crafting specific loss functions can enforce physical constraints, ensuring solutions remain true to fundamental principles. 3) Explore the use of automatic differentiation and gradient descent in training these sophisticated models. 4) See real-world applications of PINNs across diverse fields like fluid dynamics, material science, and beyond. Here is the link to register for the first lecture: https://lnkd.in/gFFUzeQb The goal of "The AI Hour" is to bring topics from various sub-domains in AI to a broader audience. We have finalized our top 6 presenters for "The AI Hour" season 1. We went through a highly selective process. We had 140+ applications of which ~40 were shortlisted first. From this list of 40, we finalized only 6 people. We were originally planning 12 episodes for Season 1. But we only shortlisted 6 people till now because we want the quality of the talk to be very high. So if we don't find more high-quality presenters, we will limit Season 1 to only 6 episodes. Enjoy the series!

89

3 Comments

I met Abhishek Gupta in 2022 and weirdly, we tried hiring each other, for our individual startup ideas. He was looking for someone technical who could help him code a restaurant PoS system. We didn't team up because we thought "ideas > cofounder" (classic mistake) Weeks later I met Abhishek, he said he would not do any intern but teach himself full stack dev so he could build whatever he wished. "That's gutsy! let's see what he does" - is what I thought 1 year later Abhishek had learnt complete end to end development and was in fact consulting other startups in software building and deploying. Then I remember at one point we were talking about B2B sales. Another year later, Abhishek had cracked into the toughest market in insurance, purely by his strategy of building great relationships with CXOs over WhatsApp. His journey was different from most people I've met. He decides, takes the risk and delivers. There's a lot to learn from Abhishek's journey of starting from scratch and building an intuition and then expertise in anything - that goes beyond the scope of this article too! So it was natural that he, Jayanth Krishnaprakash and I teamed up building Recontact, and that journey is still in the making. If you ever want good feedback on a business idea, he's a great guy to chat, given your ask good questions :)

93

1 Comment

The word "Vidya" (in NVIDIA) in Hindi translates to "knowledge" and is linked to the Hindu goddess Saraswati. Mukesh pointed out that knowledge attracts Lakshmi, the goddess of wealth. Huang responded with enthusiasm, suggesting that the name of his company might be especially appropriate given NVIDIA's growing ties with India. NVIDIA is partnering with Mukesh Ambani to build AI infrastructure in India. While I wasn’t able to attend the NVIDIA AI summit (passes sold out), it was exciting to watch the fireside chat between Jensen Huang and Mukesh Ambani. Watch the full video here: https://lnkd.in/gRauKn2w #NVIDIAAISUMMIT #AI #DATASCIENCE

19

Would you like to help a fellow student researcher on their machine learning project? When I was doing my PhD in machine learning at MIT, I used to frequently interact with passionate students who wanted a mentor on their research project. I mentored 2 such students. Both projects led to impactful publications. There are a huge number of hard working students who are passionate about AI and ML research, but don’t have access to reliable research mentors. At Vizuara, we want to make AI and ML research more equitable. It should not be restricted to just the top universities. Hard working students from all colleges should get opportunities to work on impactful ML research projects. There are so many graduate students pursuing Masters or PhD degrees in machine learning. If you are such a graduate student who wants to mentor students and guide them, please fill this form: https://lnkd.in/gZ-VpVMn You can be a co-author on these papers. These publications will help you in your graduate studies and will tremendously help passionate AI/ML students. It’s a truly win-win situation. Let us foster a community of high quality AI research, which is not just restricted to top universities.

112

3 Comments