Agenda

Abstract

This talk focuses on challenges and opportunities in designing middleware for HPC and AI (Deep/Machine Learning) workloads on modern high-end computing systems. The talk will start with an overview of the challenges in designing HPC middleware by considering support for dense multi-core CPUs, high-performance interconnects, GPUs, and emerging DPUs. Advanced designs and solutions (such as RDMA, in-network computing, GPUDirect RDMA, on-the-fly compression) to exploit novel features of these emerging technologies and their benefits in the context of MVAPICH libraries (http://mvapich.cse.ohio-state.edu) are presented. Next, the talk will focus on MPI-driven solutions for the AI (Deep/Machine Learning) domains to extract performance and scalability for popular Deep Learning frameworks, large out-of-core models, GPUs, and DPUs. MPI-driven solutions to accelerate data science applications like Dask will also be highlighted. The second part of the talk will provide an overview of the activities in the NSF-AI Institute ICICLE (https://icicle.osu.edu/) to address the challenges and opportunities in designing future high-performance edge-to-HPC/cloud middleware for AI-driven data-intensive applications (such as Digital Agriculture and Animal Ecology) over the computing continuum.

Short Bio 

DK Panda is a Professor and University Distinguished Scholar of Computer Science and Engineering at the Ohio State University. He is serving as the Director of the ICICLE NSF-AI Institute (https://icicle.ai). He has published over 500 papers. The MVAPICH MPI libraries, designed and developed by his research group (http://mvapich.cse.ohio-state.edu), are currently being used by more than 3,400 organizations worldwide (in 92 countries). More than 1.84 million downloads of this software have taken place from the project's site. This software is empowering many clusters in the TOP500 list. High-performance and scalable solutions for Deep Learning frameworks and Machine Learning applications from his group are available from https://hidl.cse.ohio-state.edu. Similarly, scalable, and high-performance solutions for Big Data and Data science frameworks are available from https://hibd.cse.ohio-state.edu. Prof. Panda is an IEEE Fellow. He is a recipient of the 2022 IEEE Charles Babbage Award and the 2024 IEEE TCPP Outstanding Service and Contributions Award. More details about Prof. Panda are available at http://www.cse.ohio-state.edu/~panda.

Abstract

Voyager, a computing resource at the San Diego Supercomputer Center, is a high-performance, innovative AI resource for conducting AI research across a wide range of science and engineering domains. The system features Intel’s Habana training (Gaudi) processors and is optimized for deep learning (DL) operations. Voyager is one of the machines with AI-focused hardware in the US National Science Foundation (NSF) resource portfolio. Voyager gives researchers the opportunity to explore unique hardware and software using well-established deep learning frameworks, like PyTorch, to implement DL techniques. The Gaudi training nodes have fully programmable Tensor Processing Cores (TPC) with tools and libraries, configurable Matrix Math Engine (GEMM), multi-stage memory hierarchy with 32 GB HBM2 memory and integrated 10X100 Gigabit Ethernet for multi-chip scale-out training. Gaudi processors were architected from inception explicitly to drive AI/ML/DL performance and efficiency, thus delivering inherent efficiency, performance and cost. 

Short Bio 

Amit Majumdar is the Director of the Data Enabled Scientific Computing (DESC) division at the San Diego Supercomputer Center and an affiliated Associate Professor position in the Department of Radiation Medicine and Applied Sciences at the University of California San Diego. His research interests are in high performance computing, computational science, cyberinfrastructure and science gateways. He is interested in convergence of HPC and data science/AI. He is PI/Co-PI on multiple research projects related to HPC and AI machines/programming, neuroscience cyberinfrastructure, and education/outreach and which are funded by NSF, NIH, DOD and industry. He is the PI of the Voyager machine and is the SDSC site PI of the Intelligent Cyberinfrastructure with Computational Learning in the Environment (ICICLE) AI Institute funded by NSF.  He received bachelor’s in electronics and telecommunication engineering from the Jadavpur University, Calcutta; master's in nuclear engineering from the Idaho State University, Pocatello; Ph.D. degree in the interdisciplinary program of nuclear engineering and scientific computing from the University of Michigan, Ann Arbor. 

Abstract

The presentation explores the foundational elements of responsible AI development, focusing on fairness, transparency, and accountability. Through practical examples and case studies, we'll examine proven frameworks for ethical AI implementation and discuss essential guardrails needed to ensure AI benefits humanity. Special attention will be given to bias mitigation, privacy protection, and governance structures that enable responsible AI deployment. As AI becomes increasingly integrated into our society, understanding these principles is crucial for creating systems that enhance rather than compromise human values.

Short Bio 

Abhijit Das is an AI researcher and senior project engineer with 3 years of experience in neurophysiological research and cognitive behavioral analysiswith expertise in developing AI solutions in healthcare. He is currently leading projects that bridge neuroscience and artificial intelligence.

Tanmay Jain is a Project Engineer at C-DAC, where he works on advanced AI solutions, including behavioral analysis using thermal imagery for forensic psychology. He completed an impactful research internship at the National University of Singapore, developing sentiment analysis models to forecast stock movements. A B.Tech graduate in Computer Science from SRM Institute of Science and Technology, he brings expertise in machine learning, AI-driven applications, and innovative software development.

Abstract

AI is revolutionizing legal eDiscovery by dramatically increasing efficiency and accuracy.  AI algorithms sift through mountains of data like emails and documents to pinpoint relevant information faster than humans, saving time and reducing the risk of missing key evidence. This technology understands language context, going beyond simple keywords, and learns from human reviewers to improve its accuracy over time. AI can even predict case outcomes by analyzing past data, empowering legal teams to make strategic decisions. Ultimately, AI automates tedious tasks, minimizes errors, and provides valuable insights, making eDiscovery faster, cheaper, and more effective.

High-profile legal cases such as antitrust cases and congressional investigations may involve tens of terabytes of data and hundreds of millions of documents in diverse formats and languages from which evidence needs to be discovered, sometimes by detecting subtle hints in phrasing or timing. We will see how AI is dramatically simplifying such difficult, overwhelming and costly projects.

Short Bio 

Sukrit Sondhi is the principal product architect for Macmillan Learning, where he provides technical leadership for innovative online learning platforms. Sukrit has nearly three decades of experience in delivering IT solutions and consulting services in education, supply chain, healthcare, government and telecommunications. 

Abstract

Success of any e-Marketplace would depend on the ensuring the “Healthy competition” and ensuring a fair playing field for all stakeholders involved. This requires 3 things – Visibility into the competitive landscape, Uniform availability of information, Reliable Benchmark. NEC’s competitive assessment suite enables these information for the stakeholders.

“Visibility into the Competitive Landscape” is enabled through the Market Intelligence Module, which focuses on enabling an unified platform to collect information from multiple online sources to aid in performing dynamic competitive assessment.

“Uniform Availability of information” is enabled through the Product Similarity Module, that standardized the data collected through various platforms to enable comparison and deriving actionable insights

“Reliable Benchmark” is enabled through the Price-Gap Analysis module, which leverages various statistical / ML methods to arrive at a benchmark for a product / category (depending on the use-case) to enable informed decision making for both sellers and buyers on the e-Marketplace.

Short Bio 

Mr. Shailen Tewathia, who is Data Sciences Manager and Specialist at NEC Corporation India, has over 9 years of industry experience, including 6 years specializing in Data Science and Data Analytics. Shailen has been at the forefront of leveraging data-driven solutions to enable decision making across various echelons of organizational hierarchy. In his current role, Shailen is working closely with various Ministries of Government of India (GoI), to bolster the digital transformation initiatives undertaken. As part of today's presentation, Shailen will be talking about "Leveraging AI for enabling Competitive Marketplace".

Ms. Abisanka Mishra, a Data Science Specialist with over 7 years of industry experience, has developed a strong expertise in Data Science and Data Analytics stream. She has played pivotal roles in utilizing data-driven insights to foster innovation and drive strategic decision-making across various organizational levels. In her current role, Abisanka is actively working with key stakeholders to accelerate digital transformation and implement impactful data-driven solutions. In today’s presentation, Abisanka will be discussing "Leveraging AI to Drive Business Growth and Innovation."

Abstract

This talk will be a reflection of the recent researches and development of a new era of robotics and AI in physical therapy interventions. The talk will focus on how the technology has advanced in recent times with the help of AI and new software developments to help increase the efficiency of robotic exoskeletons and prosthetics resulting into increased positive patient outcomes. Helping patients to achieve their pre- disorder state of movement  with the intervention of both the physical therapist and robotic systems. Creating an ecosystem and a multidimensional approach for the patient to recover from their disability. Underlying the evidence based researches and their contribution in developing the existing hardware and software used in the industry. Use of the most advanced robotic systems around the world and some projection of their user friendly interface both for the therapist and the patient . New upcoming improvements in the existing technologies to inculcate a more efficient use of exoskeletons with the patients. Talk will also include the statics and data available on the use of robotic systems in physical therapy around the globe and its effects on patient population. Finally creating a healthy approach of treatment by involving both the decision making ability of the therapist about indications and contraindications for the use of these advanced robotic systems. 

Short Bio 

Dr. Pratham Pamnani is PT (neuro), MSSI.

Abstract

Gamified learning, combined with the power of AI, is transforming education by making it interactive, personalized, and engaging. The Gamified Learning Assessment Management System (GLAMS) harnesses AI models like Llama and Gemini to dynamically generate customized assessments, enabling educators to create adaptive learning experiences effortlessly. Through intelligent matchmaking and AI-driven bots, GLAMS ensures seamless participation, balanced competition, and continuous learning for all students. By integrating game mechanics with advanced analytics, the platform motivates learners and empowers educators to tailor instruction to individual needs, unlocking the potential for inclusive, impactful education. With successful implementation at Army Public School, Danapur, BRC Patna, and ISEA, MeitY GLAMS is a testament to how gamification and AI can reshape learning.

Short Bio 

Amarjeet Sharma is a seasoned Software Engineer with over 11 years of experience designing, developing, and maintaining robust software applications. His professional journey has been marked by significant contributions to research and development in areas such as Parallel Programming, Software Development, High-Performance Computing (HPC), Artificial Intelligence (AI), Data Science, and HPC System Administration. With a strong foundation in technology and a proven ability to collaborate with industry stakeholders, a dedication to excellence, and a forward-thinking approach to creating high societal value projects.

Aniket Pattiwar is a dynamic full-stack developer and AI technology innovator with expertise in web technologies and educational innovation. With two years of experience, they have successfully developed projects across the MERN stack, React.js, and AI platforms, notably presenting at the Global Partnership on Artificial Intelligence (GPAI) Summit 2023. He excels at creating adaptive, impactful technological platforms that bridge software development and educational transformation.

Abstract

One key to achieving an organization’s mission is finding ways to measure its performance and improving on them. Operational metrics tracking shows how well the organization performs its mission. At the National Energy Research Scientific Computing Center (NERSC) we have developed innovative techniques to automatically collect data and generate and distribute performance metrics reports to the leadership team at NERSC and beyond. The metrics are essential for our leadership team to make data driven decisions.

Short Bio 

Kadidia Konaté is a data scientist at the Lawrence Berkeley National Laboratory (Berkeley Lab) in the National Energy Research Scientific Computing Center (NERSC). She leads efforts in machine learning for high performance computing problems, including workload characterization and anomaly detection across data centers. She is currently pursuing her MBA at the University Of California Haas School of Business. She received an Eng. Degree (M. Eng) from Ecole Centrale Paris, a top French engineering school (Grandes Ecoles) in Paris, France in 2013. She was Summa Cum Laude at the Fourier Institute in Grenoble, where she received her B.S. degree in fundamental mathematics in 2010.

Abstract

TBD

Short Bio 

TBD

Abstract

TBD

Short Bio 

TBD