Is there an algorithm that predicts the outcome of sindhi genes in a machine learning lab? Dr. Naresh Manwani brings his unique qualities into his research in Responsible AI, building ML generative models and exploring applications of AI and ML in NLP, CV and Computational Neuroscience. He is particularly excited about the Institute’s work in Causality.

An Assistant Professor at IIIT Hyderabad’s Kohli Centre on Intelligent Systems (KCIS), Manwani’s diverse research areas are in AI robustness, explainability, fairness and privacy, learning with weak supervision and in high risk situations, adversarial machine learning, generative models and applications, automating exploratory data analysis (EDA), Causality as well as applications of AI in NLP, CV and Computational Neuroscience. He brings over four years of industry experience at Microsoft Research India, Microsoft India (R & D) and GE Global Research to his work in the KCIS Machine Learning Lab.

Touchdown at IIIT Hyderabad
“I come from a hardcore Sindhi business family in Ajmer. A Ph. D in my community was quite uncommon”, laughs Manwani who completed his PhD from IISc Bangalore and interned at Bangalore’s Microsoft Research Lab. He worked at GE Global research center Bangalore, on various non-conventional applications of AI in steam turbines and fault prediction in aircraft engines. He was recipient of the above and beyond awards for his contributions to certain projects.  To explore the ecommerce space, he moved to the Bing ads platform at Microsoft R&D Lab. While his product-driven research in AI and ML was exciting, he realized that his heart was in academic rigor.

“IIIT Hyderabad has the biggest AI hub and in a single building there are almost 50-60 faculty, working in AI and related areas. That’s when I decided to move to IIIT Hyderabad in 2016”, says Manwani who teaches topics in Probability and Statistics, Linear Algebra, Optimization Methods, Topics in ML, Deep Learning Theory and Practices at KCIS.

Exciting work at the Machine Learning Lab
A lot of his research work today centers around robustness, ethics and fairness, interpretability and privacy; building ML algorithms to handle complexities, with human-like virtues in critical decision making. His study of decision-making in very high risk situations has huge applications in domains like judiciary, finance and medical diagnostics. Recently, his students developed a theoretical framework for calibrated losses for adversarial robust reject option classification that was Runner up for the Best student paper in ACML 2024.

“For a Bosch collaboration, we introduced robustness feature into an autonomous driving tool that detects oncoming objects with very good efficiency. Recently, our dual-degree student Rishabh Bhattacharya who is working on the object detection segment of the project, was able to resolve a similar task in the surveillance domain for the Naval Hackathon 2024”. The theme was Navigation and real-time tracking of flying objects, for which he received the top prize from Defense Minister Shri Rajnath Singh Ji.

Learning with weakly supervised data is another key research topic of interest for him. “With Dr. Vineet Gandhi, we have developed a self-supervised learning algorithm which beats all the state-of-the-art methods in its performance.” With IBM research, his team developed an algorithm for automating the job of a data scientist; working on auto exploratory data analysis (EDA). This work won Research Track Best Paper Runner-up award at the AIML Systems 2024 conference. “For a joint project with JPMC and Dr. Sujit Gujar on federated learning, we tried to build an efficient AI-ML model by collating data from disparate parties while ensuring privacy guarantees. Our paper is currently under review in a conference”, reports Manwani who is part of two DST-funded core research projects. For a startup, in collaboration with IIITH Prof. Praveen Paruchuri and a senior faculty in UT Austin, Dr. Naresh developed a tool to detect malicious activities in the stock exchange.

About upcoming research that excites him, Manwani states, “Causality is a desirable aspect in many AI applications and I hope to take some courses in causality theory or causal inferences”.

IIITH campus
Campus is the best part of academic life, believes the scholar. “I feel connected with everyone on campus and know them by name! Faculties and students from across India celebrate all festivals together, in the true spirit of diversity. Being a primarily undergraduate institute, students are very energetic and you can feel the vibrations, whether you are teaching, working or interacting with them at cultural festivals or the myriad events happening on campus”, says Manwani, once an ace debater at school and district level. “This is a lively place to be amongst youngsters who not only challenge your research but your fitness levels. We recently formed a group to train for marathons”, he informs.

Manwani has an interesting research portfolio that includes over twenty-five Talks and posters, 35 conference and workshop papers and six Journal publications and a book chapter. Manwani has served as program committee member in several major conferences like ICML, AAAI, AIStats, CVPR, ICLR, AAAI, IJCAI, ACML, PAKDD etc.

A brainy cocktail -Sindhi business genes and scientific temper
“Growing up in Ajmer, we came from large families, where everyone joined the family business at a very early age. I was 12 years old when I started helping my father at work.  “My parents always gave me the freedom to take decisions of my life and I am very thankful to them for the nurturing environment that they gave me and my three siblings”.

Manwani finished his schooling from the local Gujarati Senior Secondary School. He credits his school science and maths teacher Mr. Ramlal Saraswat for pushing him to be a better person and student. “He had a lot of faith in me and that was a turning point in my life”. He credits his PhD advisor IISc. Prof. P.S. Sastry for shaping him to be an independent researcher.

An occasional reader, Manwani prefers popular authors like Chetan Bhagat and Paulo Coelho and has recently started reading Swami Mukundananda. “My wife Loveleena and I watch two-three movies together every week but we are very particular about what we watch”. They have a four-year-old son Reyansh who considers the campus as his playground and extended home. “Everyone, from the security guards to faculty members are so warm and welcoming in their interactions with him”.

“My philosophy in life is to enjoy every moment. For the last decade, I have been following various fitness activities like yoga, running and strength training” says Manwani who has completed a couple of half-marathons and is prepping for his first full marathon next year.  “A life of academic research gives you good flexibility to manage time as per your priorities. I have a fixed scholastic schedule to teach classes but apart from that, I have the flexibility to run errands, manage my work and family life. I get my inspiration from nature. I am grateful for living in this invigorating and enriching research environment. Every morning when I wake up, I know that there is something significant that I am going to be working in”.