From the lab to the limelight 🧠✨ The Conclave on AI Governance by WSAI & CeRAI, IIT Madras continues to make headlines — spotlighting conversations on regulation and responsible, inclusive & participatory AI. This conclave was an official Pre-Summit Event for the India-AI Impact Summit 2026 (https://lnkd.in/gwwdFEkc) 🔮 The future of AI Governance begins here. #IndiaAIImpactSummit2026 #IndiaAI #CeRAI #WSAI #IITMadras #ResponsibleAI #AI Balaraman Ravindran | IndiaAI | Centre For Responsible AI (CeRAI) | Robert Bosch Centre for Data Science and Artificial Intelligence (RBCDSAI) | IBSE: Centre for Integrative Biology and Systems medicinE | AI4Bhārat | Walmart Center for Tech Excellence | Indian Institute of Technology, Madras | Ministry of Education, Government of India, New Delhi

𝐈𝐧𝐝𝐢𝐚 𝐒𝐞𝐥𝐞𝐜𝐭𝐬 8 𝐅𝐢𝐫𝐦𝐬, 𝐈𝐧𝐜𝐥𝐮𝐝𝐢𝐧𝐠 𝐈𝐈𝐓 𝐁𝐨𝐦𝐛𝐚𝐲, 𝐭𝐨 𝐁𝐮𝐢𝐥𝐝 1-𝐓𝐫𝐢𝐥𝐥𝐢𝐨𝐧 𝐏𝐚𝐫𝐚𝐦𝐞𝐭𝐞𝐫 𝐀𝐈 𝐌𝐨𝐝𝐞𝐥. 𝐀 𝐁𝐨𝐥𝐝 𝐒𝐭𝐞𝐩 𝐢𝐧 𝐀𝐈 𝐋𝐞𝐚𝐝𝐞𝐫𝐬𝐡𝐢𝐩. 👉 India has announced a major initiative to develop a massive 1-trillion parameter artificial intelligence model. 👉Eight leading organizations, including IIT Bombay, have been chosen to spearhead this ambitious project. 👉The move positions India among global leaders in large-scale AI development. #IndiaAI #TrillionParameterModel #IITBombay #AIInnovation #TechForGood #MachineLearning #DataScience #AIResearch #FutureOfAI #DigitalTransformation https://lnkd.in/dYc6V4-t

We are thrilled to share that two research works from Prof. Tushar Shinde’s group at IIT Madras Zanzibar have been accepted at the prestigious Women in Machine Learning (#WiML) workshop, #NeurIPS 2025 Efficient Spam Detection with Sentence-BERT using Adaptive Uncertainty-Diversity Ranking Coresets – by Aisha Hamad Hassan & Tushar Shinde From National Goals to Industry Action: AI-Driven Forecasting of India’s Carbon Emissions – by Anushka Chaubey & Tushar Shinde Both works received reviews as the Top 15% of accepted papers! Kudos to our brilliant Master’s students, Aisha and Anushka, for their outstanding contributions. #NeurIPS2025 #WiML #MachineLearning #AI #ResearchExcellence #IITMadrasZanzibar #WomenInTech #AIResearch

We are pleased to share the journey and reflections from Cohort 7 of the Capacity Building Program on Artificial Intelligence (CBPAI), conducted by Pavani Ayinampudi, PhD, Ed.M & Shreya Singhal — Ed.M. Harvard alumni. This cohort was unique, being independently designed and delivered, and the video highlights authentic participant feedback, valuable insights, and the impact created. 🌟 Their efforts showcase how faculty across India are reimagining teaching and learning with AI, and how such initiatives continue to strengthen the CBPAI program. 🚀 #CBPAI #FacultyDevelopment #HarvardAlumni #IITRopar #EducationDesign #AIinEducation

Hopefully, indigenous AI models will soon be developed. However, one of the major challenges for Indian researchers, scientists, and industries is the lack of India-specific data to train these models. The Government of India and academic institutions must take the lead in creating open data repositories tailored to the Indian context. #IndiaAI #ArtificialIntelligence #IIT #IIIT #NIT #Research #MachineLearning #DeepLearning #Data

#AAAI2026 #IAAI2026 Innovative Applications of AI Theoretically sound ML pipeline: You have a large dataset. Split it into train validation and test set. Never touch the test set. Develop your ML method on train set. Tune hyper parameters with validation set. Report your results on test set. Do this independently on 100 datasets. Great paper accept in top ML conference. Problem: Fails miserably in practice. Take a well established model on one dataset and apply to a different dataset for the same problem. A simple coin toss becomes better than a “well designed” ML method. Reason: Theory claims to capture generalization with one big assumption. The training data sufficiently represents statistical properties of the data representative of the problem. Solution: Get rid of the mindset that AI/ML can take over the world. Recognize that the knowledge that humans have gathered collectively through multiple civilizations cannot be replicated by AI. In our recent paper on “Experience with Single Domain Generalization in Real World Medical Imaging Deployments” in AAAI 2026 innovative applications of AI we show that if we incorporate doctors knowledge into our ML pipeline it improves performance and explainability. Thanks to #NIH and #NSF for partially funding this research. The work also highlights the challenges we are facing at EpiPrecision AI Sandeep K. S. Gupta Komandoor Srivathsan M.D. #SCAI #ASU #medicalimaging #MayoClinic #IMPACTLAB_ASU

Pleasantly surprised by perceptive positive reviews from #IAAI2026 giving validation to years of hard labor by many humans from the IMPACT Lab at ASU and crucial domain insights from experts at Phoenix Children and Mayo Clinic. As a newly minted #Noble Laureate in Economics Joel Mokyr said in a recent talk, AI is just a tool not an existential threat (at least not in the foreseeable future despite what last year's Laureate in Physics might have said regarding AI) but a capable (research) assistant not much unlike a telescope or a word processor. In this work we show how to carefully craft such an AI assistant for #medical #clinical tasks in Epileptic zone localization, Heart Arterial Stenosis prediction, and Diabetes Retinopathy grading - not much unlike the task of carefully crafting a telescope for clear distant vision - e.g hours of polishing lenses. Yes it's going to be hard work involving manual labor and ingenuity to make these AI tools so that they faithfully serve us humans! If AI is the new electricity then we need to carefully think about how to make it safe to let it enter in our homes, workplaces, hospitals etc. - if you have not lately, go and look into the fuse box in your home electric supply unit, think about the electric voltage that's permitted and all the safeguards that are put in place to make sure accidental misuse doesn't cause irreparable damage and fatalities. To borrow from Prof. Mokyr - like economics, AI is entering an era where empirical work will enjoy greater relevance as opposed to theoretical or bench work - so that its.promises can be fully realized in our (my) lifetime. We are more than glad to talk to anyone who cares about nitty gritty details - trials and tribulations - of deploying AI that is ready for the real world! #PCH #MayoClinic #Impactlabasu #AIforRealWorld #medicalAI #ClinicalAI #safeAI #epilepsy #CADprediction #DiabeteseRetinopathy #ASU #FSE #SCAI

🚀 Successfully attended a highly insightful session on “AI in Engineering Applications, Innovations and Future Trends” held on 16th September 2025, organized by the Pakistan Engineering Council (PEC). The session provided valuable knowledge on how Artificial Intelligence is reshaping engineering practices, driving innovation, and opening new possibilities across multiple domains. Key discussions revolved around: ✅ Practical applications of AI in engineering workflows ✅ Emerging innovations transforming industries ✅ Future trends that will redefine engineering practices ✅ Opportunities and challenges in adopting AI for sustainable development Grateful to Pakistan Engineering Council for arranging such an engaging and forward-looking session that broadens our vision toward the future of engineering with AI. 🌍🤖 #ArtificialIntelligence #Engineering #Innovation #FutureTrends #PakistanEngineeringCouncil #LearningAndGrowth #AI

I had the opportunity to present my paper “A Novel Approach in Feature Selection using Ensemble Technique for Predicting Important Features in Datasets” at the International Conference on Next Generation Technologies for Sustainable Computing (ICNGT-2025), organized by Shrimati Indira Gandhi College, Tiruchirappalli. This work introduces a new ensemble-based feature selection method that integrates CFS, RFE, and LFS, improving accuracy in high-dimensional datasets using SVM, Decision Tree, and Random Forest classifiers. 📌 Key Highlights: Reduced irrelevant features for faster & more accurate ML models Outperformed traditional feature selection methods Achieved higher accuracy across benchmark datasets I’m grateful to my guide Dr. P. Lakshmi and co-author Sree Kaviya V. for their collaboration and support. Looking forward to applying this research in real-world AI & ML applications and connecting with professionals in the field. 🚀 #MachineLearning #AI #FeatureSelection #Research #EnsembleLearning #Conference2025

AEC Hosts ATAL VAANI Conference on AI & ML Asansol Engineering College successfully inaugurated the ATAL VAANI Conference on Artificial Intelligence and Machine Learning in collaboration with IQAC and the Department of IT. The 2-day conference, held on 18th & 19th September 2025, brought together academicians, researchers, and industry experts to explore the transformative role of AI in bridging new relationships between humans and machines. The event was inaugurated with the ceremonial lamp lighting by: Dr. Gourishankar Panda (Vice Principal, AEC) Dr. Samarjeet Kar (NIT Durgapur) Mr. T. T. Lahiri (SAIL) The conference featured more than 30 research paper presentations, providing a platform for scholars and students to exchange knowledge and ideas. A special publication, ATAL VAANI Souvenir, was also released during the inaugural ceremony. This proud moment for AEC was published today in Bartaman Patrika (National Daily, Asansol Edition), further highlighting the institution’s contribution to research and innovation. www.aecwb.edu.in #AEC #ATALVAANI #AI #MachineLearning #Research #Innovation #IQAC #Conference2025 #BartamanPatrika