pannalas

The Dawn of a New Technological Era: Riding Four Exponentials at Once

AI/ML, Biology and Life, Computing, Energy, Innovation, Science, Technology /

It’s an incredibly exciting time to be someone who works across different fields of science and tech. With all the doom and gloom, we hear about R&D policy and geopolitical uncertainties, it’s easy to get pessimistic. These are the times when one needs to put on a dreamer’s hat and there’s actually immense cause for optimism. We’re at a really unique point in history, where four powerful exponential trends are all coming together (https://ekta.net/LI/ExpConv.html). For what is possibly the first time, the foundational pillars of innovation are all accelerating simultaneously: Raw computational power at your tips (exaflops, vibrant open-source software and tools ecosystem, tremendous algorithmic developments including AI/ML, reduction in costs for everything data – storage, memory, bandwidth) Abundant, sustainable, and cheap energy (electrons from solar and storage; associated power electronics and controls; energy efficient devices such as lights, cameras, heat pumps) The deep wisdom of our planet’s four-billion-year evolutionary history (genomics, evolution of intelligence and life, biotech, protein folding, medicine and microbiome) These aren’t just small steps forward, but these tectonic shifts will overcome the noise we are seeing today. It’s going to be a total game-changer in what we can do to create real value for society. The big names of the last tech boom, like Amazon, Google, Apple, Meta, Intel, Microsoft, Nvidia, and AMD rode the powerful wave of Moore’s Law. Now, we have a chance to ride not just one, but multiple, intersecting “Wright’s Laws” at the nexus of the digital and physical worlds. The things we can do now by democratizing power directly from sun while leveraging digital tools and evolution were just science fiction a decade ago. Just think about what this means. Imagine having the sensory ability of a dog’s nose on your watch, a device that could continuously check for the faint hormonal signals of stress or the early biomarkers of disease. This would totally change personal medicine and health. Or think about shoes that can pick up subtle vibrations like animals do, giving you feedback on your run but also alerting you to an approaching tsunami, earthquake, or tornado. This convergence lets us tackle big, systemic problems in a whole new way. We can now aim to create truly explainable AI models of financial markets, letting us make investment or policy decisions based more on numbers and probabilities instead of just opaque, black-box predictions. We can engineer solutions like the “Shunya” (zero-impact) toilet (https://ekta.net/LI/shunya-waterless-toilet-v1.html), a standalone unit that, powered only by the sun, converts human refuse back to its basic, harmless components: CO₂, H₂O, and valuable minerals. Or we can build smart traffic control systems that see a city’s traffic as a single, fluid system, optimizing flow to cut down on congestion and even suggesting when you should leave home or work based on your schedule while optimizing the whole system. The possibilities are endless when we use these converging technologies to create sustainable solutions that improve our quality of life. It’s not a question of whether we can solve these big challenges anymore, but how we’re going to do it? Put on your dreamer’s hat and list out what else becomes possible at this convergence?  

The Dawn of a New Technological Era: Riding Four Exponentials at Once Read More »

Simple Introduction to the Geometry of Learning

AI/ML, Biology and Life, Computing /

Everyone learns differently. For me, the most intuitive way to grasp complex ideas has always been through geometry and visualization. This is how I learned machine learning techniques like LLE (Local Linear Embedding) and ISOMAP when I was first introduced to them in 2007. Long before that I was thinking about the reduction of big data to small data through various multiscale techniques in the pursuit of constructing reduced order models. When I see the shape of a problem, I begin to understand its structure. This is the simple idea behind a mini project I’m excited to share: an interactive webpage designed to explore the “Geometry of Learning.” My goal was to take some of the most abstract and powerful concepts in modern AI, ideas like high-dimensional data, manifolds, transfer learning, and latent space, and make them more tangible and accessible. I wanted to simplify these topics so that more of us can understand and use the AI tools effectively, particularly Large Language Models (LLMs), that are rapidly becoming a part of our daily lives. Whether we realize it or not, LLMs now underline many of the technologies we use, from creative tools to a simple Google search. As these algorithms become more integrated into our world, it’s crucial that we develop an awareness of how they work. Understanding that an LLM learns by creating a compressed, geometric “map” of reality (a latent manifold) helps us appreciate both its incredible power and its inherent weaknesses. This knowledge is our best defense against falling into the “hallucination trap,” those moments when an AI produces a confident, plausible-sounding answer that is completely wrong. By understanding that the model is navigating a learned map, we can better appreciate when it’s on a well-trodden path versus when it’s wandering into uncharted territory where errors are likely to occur. I’ve also added a forward-looking section to the webpage. My goal here is to think about how we can build on the massive investments already made in Generative AI to advance the accuracy of predictions and reduce hallucinations. It’s my belief that the current path of simply scaling up models might not be enough to get us to Artificial General Intelligence (AGI). Real breakthroughs will be necessary, and I hope this exploration gives some context for where we might look for them. I hope you find this webpage helpful. I’ve tried to keep the content as pedantically correct as possible while simplifying it for a broader population. A quick note on the animations you’ll see: they are conceptual tools designed to illustrate the underlying principles. They are not meant to be exact reproductions of the complex reality of these systems, but rather visual aids to help build intuition. Please explore, interact, and I hope it helps you see the beautiful geometry that underpins modern AI. I have some foundational references at the end for you to explore this topic in-depth.

Simple Introduction to the Geometry of Learning Read More »

Ideas Are Funny Little Things

AI/ML, Biology and Life, Computing, Innovation, Science, Technology /

“Ideas are funny little things. They won’t work unless we do.” This is the quote on a plaque in my living room. It’s the only heirloom I valued enough to bring from my childhood home in India. As you can see from the image, the plaque has gained round edges and chips from all the falls it has endured over four decades. My own learnings, shaped by the computational transformation of our lives, has been similarly molded. I feel the need to create a new plaque that reads: “Ideas are funny little things. They won’t work unless we leverage machines and do.” I was a curious kid with a continuous barrage of questions: Why? How? What? This desire to know, however, was somewhat suppressed. I grew up in a society that encourages children to speak only when spoken to, and in a middle-class family, there weren’t many who had the time to entertain a child’s never-ending questions. This steered me toward becoming a more introverted boy who learned by observing, creating mental frameworks to make sense of the world, and developing systems thinking to connect things on multiple levels. This inward journey of hypothesizing and justifying outcomes became natural, though sometimes in ways that weren’t grounded in reality. In a way, I was like a modern LLM, living in my own world of hallucinations. I was a curious explorer with a passion for cricket until I changed schools after the 8th grade. It was then that I became serious about studying for the engineering entrance exams, and I was fortunate enough to be accepted into an IIT. Throughout my undergraduate and graduate days, most of the coursework bored me. The one thread that continuously inspired me, however, was our growing ability to model and simulate the world around us. When you model something from first principles, you can answer questions like “Why?”, “How?”, and “What?” and perform all kinds of what-if analyses. I could finally build on my childhood explorations, but now they were grounded in solid scientific and theoretical underpinnings. Fate would have it that I ended up working on one of the most challenging computational topics: the turbulent combustion of two-phase flows, a field at the intersection of fluid dynamics and chemistry. It was a humbling experience. I could only make a small dent in the field, and there has been only limited progress in the 30 years since. However, this journey raised new questions in my mind. By working across the disciplines of domain sciences, applied math, computer science, and statistics, I developed a deep appreciation for what I call computational thinking. This is the ability to take a problem and convert it into a computational one that can be solved in-silico. This conversion requires a blend of skills: domain knowledge (physics, chemistry, biology), the ability to map that knowledge onto theories and equations with appropriate approximations (engineering), an underlying understanding of the algorithms to solve those equations (applied math), the knowledge of how to best solve it using massive parallelization (computer science), and robust analysis (probability and statistics). It’s also critical to ground this process in the specific question you’re trying to answer, the level of fidelity needed, and the constraints of time and money. By tackling one of the toughest computational problems, coupled with access to the largest supercomputers of the time and a background in math, CS, and physics, I started developing the skills for a career in computational science. I was recruited from Georgia Tech while still finishing my Ph.D. to help a national lab respond to an important DOE initiative called SSI (Strategic Simulation Initiative), which aimed to develop a high-fidelity digital twin of an Internal Combustion Engine. I visited Sandia National Laboratories in Livermore even before officially joining Oak Ridge National Laboratory (ORNL) to help lay down a stake for the lab in this $70 million initiative. In a twist of fate, the project’s funding was eliminated by Congress. In retrospect, this was fortunate for me, as it forced me out of my comfort zone to explore applications beyond my niche. The bifurcation chart below shows my evolution from those early days at Georgia Tech to a Distinguished Staff Member at ORNL, and now, for the last 10 years, as a Research Fellow at SABIC. At ORNL, I was a true explorer, working on a broad range of problems and across disciplines with some of the best minds in computational science. Later, I was recruited by SABIC to focus on a narrow set of capabilities I had developed at ORNL. Working in a corporate research organization helped me focus on a few key problems and deliver solutions. We developed and optimized a novel reactor entirely in-silico, de-risking it through extensive experimentation. If successful, it will displace the century-old steam cracker for converting hydrocarbons to olefins. We also developed unique solutions that leverage the low melting point of plastics to create self-quenching, safer batteries. Through work with the World Economic Forum and other petrochemical companies, I also gained exposure to policy and the need for broader stakeholder engagement to bring sustainability solutions to market. Over the last three decades, many ideas have been brewing in my mind. I never had a chance to develop them further because I was limited by my own abilities or because they were adjacent to my core expertise, requiring dedicated time I didn’t have. Time is a finite resource, and it was difficult to translate those ideas into proofs-of-concept (POCs) and actionable plans. Fast forward to today. I am having a blast turning many of those lurking ideas into POCs and actionable plans. This is possible thanks to the computational thinking I’ve honed over the last 30 years and my curious explorations over five decades. What made the difference? LLMs, AI/ML tools, the open-source software ecosystem, and low-cost hardware have dramatically reduced the barrier to fleshing out ideas and creating prototypes. It is an exciting feeling to pour life into one’s ideas, weed out the bad

Ideas Are Funny Little Things Read More »

Heterogeneities, tyranny of scales, and curse of dimensionality

AI/ML, Computing, Energy, Innovation, Materials, Science, Technology /

In most complex physical systems (and to some extent others like financial markets), competing non-linear forces lead to instabilities and organized large scale structures. The overall dynamics are controlled by the large scale structures to a great extent but those large structures cannot be accessed directly as they are the result of the small scale interactions.  In CFD (Computational Fluid Dynamics), we are stuck in the pursuit of conducting even higher resolving simulations to get to the ground truth by modeling all those small scale interactions that lead to the emergent large scale structures. We are creative in coming up with subgrid models to reduce the computational load to resolve absolutely what we have to and have correlations for the unresolved scales. There are other heuristics that we apply to things like turbulence and drag closures, etc. to make the problems computationally tractable.  I have been introduced to Machine Learning by Badri Asokan in 2007 and the objective then was to find low dimensional manifolds, segment them into linear spaces, and map PODs (Proper Orthogonal Decomposition – a derivative of PCA) onto those linear spaces with the objective of developing reduced order models. We made some progress on a toy problem (spouted bed) but the proposal to generalize and further this idea was unfortunately not funded and we never got to develop this further. On the other end of the spectrum, I worked with Stuart Daw and Jack Halow to work on an agent based model in 2003 to simulate bubbling fluidized beds with bubbles with simple interaction rules. As I shared recently, these models run in real time and can predict emerging behavior such as slugging quite well but are limited in their own ways. Any extensions that I tried to do to make them more attuned to data failed.  Having a taste of methods to unravel low dimensional manifolds and also realizing that most systems we are trying to model lie in much lower dimensions than than the millions or now billions of degrees we bring in through CFD, I presented concepts at various venues on how we can use AI/ML to break these vicious dependencies between heterogeneities, tyranny of scales, and curse of ever increasing mesh sizes. Now it is more imperative to revisit as the LLM boom brought us into our laps the neural networks based learning algorithms that give access to low dimensional latent spaces and possibly transfer learning from similar phenomena, hardware co-designed with software for deep learning on large datasets, and plethora of software and data analysis tools available to explore all aspects of heterogeneous structures in complex systems. Presenting a concept without actual data or a proof-of-concept doesn’t catch the attention as I was hoping someone smarter than me will pick it up and pursue those concepts. Traditional CFD codes (including the many I have developed or contributed) are not flexible to try out new concepts and I never could do this on my own till recently without significant investment of my time.  With the advent of LLM tools, it has become easier to try out new ideas and also create a proof-of-concept that one can interact with live to understand and later utilize this understanding to advance methods to model systems with a computational dimensionality that matches the underlying physical dimensionality.  Modeling risers in fluid catalytic crackers (FCCs) turned out to be one of the most difficult problems of great practical relevance emblematic of the topic of this post. Particles in the riser collide and dissipate to form clusters and these clusters with reduced drag drop down to create interesting phenomena like core annulus, create back mixing, high solids concentration at the bottom of the riser, and control the solids holdup and gas-solids contacting. We made tremendous progress on this front with the filtered-drag and EMMS models. One thing I wanted to explore was to reconstruct the large scale structures and use that to modify the drag as based on my experience with modeling a square cross-section circulating fluidized bed (particularly in 2D), the drag reduction from clustering is key to the solids fraction profile along the axis. The interactive website (https://ekta.net/LI/Riser-DHRDM.html) will introduce Dynamic Heterogeneity-Resolving Drag Model (DHRDM) – a dynamic way to detect clusters based on some heuristics that you can play with and that will affect the solids holdup and distribution along the axis. This is a grossly simplified version to play around and develop intuition. There is a lot more work to be done in terms of automating all this by training AI on large datasets where one can identify these structures based on local parameters and use more sophisticated corrections for drag based on the cluster shape. It is exciting that the modern AI tools help us to play around with our ideas – from start to finish, I must have spent around 5-6 hours and most of that time is to get the plots right.

Heterogeneities, tyranny of scales, and curse of dimensionality Read More »

We’ve Been Solving the Wrong Problem – introducing The Shunya Toilet

1 Comment / Uncategorized /

Some problems aren’t just problems; they’re personal. For me, that problem crystallized at a public event in Knoxville, Tennessee. I took my two-year-old daughter to a portable potty, and the moment I opened the door, she recoiled. That look of pure disgust on her face was a gut punch. It sent me right back to my childhood in India. I grew up seeing the reality of the sanitation crisis. Summer visits to my village meant open fields were the only option. In the cities, you saw the state of public restrooms, the rampant public urination. That porta-potty in Knoxville was just a chemical-laden, plastic version of the same fundamental failure. We haven’t solved the problem; we’ve just found better ways to hide it. This is the contradiction that drove me: we have the technology for space travel with ambitions to go to Mars, we have 85″ LED TVs for less that $1000, yet nearly half the world lives with a daily, undignified, and dangerous reality when it comes to sanitation. As an engineer, I’ve dedicated my career to energy and thermochemical conversion. I look at a problem and see systems. I see inputs and outputs. And the current sanitation equation is broken. To be clean, it needs water. To be deployed globally, it must be waterless. To be safe, it can’t use chemicals. To be sustainable, it can’t create waste. To be mass marketable, it has to be cost effective. For too long, we’ve accepted these as unavoidable trade-offs. It’s time to stop. The answer comes from a new convergence. For the first time, the tools we need are all within reach: low-cost solar, efficient LFP batteries, and the incredible power of AI to automate and control complex systems. We can finally build something that doesn’t compromise. That is why I am proud to introduce the concept of The Shunya Toilet. “Shunya,” in Sanskrit, means zero. It’s a concept that came from India, representing not just a number, but a new beginning. It is the foundation of our design. This isn’t just another toilet. It’s a fundamental rethinking of the entire system. It’s a closed loop that respects both the user and the environment, inspired by the way nature operates. In nature, nothing is waste; everything is reused. We are simply using our advanced knowledge of thermochemical conversion to accelerate that natural cycle, using high temperatures to do in minutes what takes nature years. The full technical details are something I look forward to sharing soon. For now, I wanted to share the vision, the promise of starting at zero to finally solve one of humanity’s oldest and most pressing challenges. Experience the Interactive Shunya Toilet Demonstration Here The time for compromise is over. The time for Shunya is now. Sreekanth Pannala Inventor, The Shunya Toilet For additional details, please email: [email protected]

We’ve Been Solving the Wrong Problem – introducing The Shunya Toilet Read More »

Where was I on the exponential computing journey?

Uncategorized /

I just got a 90-teraflop (AMD Ryzen AI 7 with NPU and Nvidia 5070 RTX GPU) laptop for $1200, and that made me reflect on the history of computing. This is like having a supercomputer worth tens of millions of dollars and consuming ~10 MW of power in 2005, right on my kitchen desk and completely dedicated to me, 24/7. The exponential increase in computing has been a constant over the last seven decades, and it is only accelerating with the recent emphasis on AI infrastructure. All of us have benefited in diverse ways from this rapid advancement. Below is my attempt to summarize the progress in hardware and the dramatic reduction in cost. I have superimposed important milestones from my own career where I was able to harness that wave. This has transformed my work in many positive ways, and the trend is only going to accelerate, requiring everyone to think computationally to navigate this never-ending current. In the interactive chart below, you can add a milestone of significance to you and see where it fits within this exponential growth. Here is some more background on my milestones: To contextualize this history, we can plot specific systems from a professional’s career onto this broader performance curve. This exercise transforms an abstract historical trend into a tangible personal journey through the evolution of computing. 1992: The First Rung on the Ladder 1994: Diving into High-Performance Computing 1995-1999: Riding the Teraflop Wave 1999-2015: At the Forefront with ORNL Below is a screenshot of the interactive webpage, and here is the link for you to try it out: https://ekta.net/LI/exponential-compute.html

Where was I on the exponential computing journey? Read More »

Dawn of a golden era for computational science!

AI/ML, Biology and Life, Computing, Energy, Innovation, Materials, Science, Technology /

For those of you who are not familiar with computational science – it is the discipline at the intersection of domain science, computer science, applied mathematics, and engineering. In a nutshell, it is the link between everything you touch and feel (natural and human-made) and rapidly growing computing. The most challenging science and technology problems whether related to life sciences or physical sciences fall into the grand challenge category in the simple 2×2 classification of AI Application Landscape Heatmap below and posted on LinkedIn. These problems are hard and yet to be solved as the problem dimensionality (one can think of number of parameters to explore all the solution space) is extremely high and in turn would need a lot of experimentation and data. AI, machine learning, and good statistical science can help in reducing experiments, but they cannot create more information than what is contained in the data. This is where computational science plays an important role in bridging domain science knowledge with AI/ML tools. If you think of an equation, theory, or laws (E = mc2; Newton’s laws of motion or theory of general relativity; laws of thermodynamics), they are the compact representations of our knowledge of the universe and underlying processes. Blackbox models like AI/ML or many statistical tools throw away all this knowledge and relearn all the basic interactions as they are trained on available data. This is where for these grand challenge problems it is important to incorporate all available knowledge into efficient processes where only new things are learnt and thus reducing the dimensionality of the grand challenge problems. This in turn would make solutions to these problems more accessible to AI and ML. If we look past all the debate of the value of the AI investments or whether AI will become sentient and reach human intelligence or superintelligence, what is fundamentally happening is enormous investments in computing hardware with unprecedented explosion of computing power. The hardware (computing, memory bandwidth, storage, wired and wireless networking) is getting exponentially cheaper by the day and more power efficient. This is leading to ubiquitous computing where we have powerful computing devices in our hand (a supercomputer not so long ago) and connected to the backend mega servers at lower costs than ever before. This trend is only going to continue as mentioned in the recent AMD’s Advancing AI Event (https://youtu.be/zCOqJx3Yst4?si=X_pPgYr3WGJVKnMY); the expected investments in AI are going to cross $500B by 2030 and currently the market is projected to have an incredible CAGR of 80%. If I translate that to increase in computational power, that is greater than 80% rate as computing is getting lot cheaper by the day. This means that this is the 2007 internet moment for computational science. The hardware is coming, and the question is how can we leverage decades of efforts in computational science in building the knowledge, the tools, the algorithms, the connections between computing and real world, and leverage advances in AI/ML and the associated ecosystem to rapidly create ecosystems to create value to the society via both natural and physical sciences. This will span innovations in health sciences, energy, materials, weather and climate, sustainability, defense, and in short across all sectors. This also means that computational thinking is an important skill for the modern workforce – the ability to translate problems and processes into streamlined instructions to the ever-increasing computing that is going to be all around us. As I have posted recently, AI is already enabling rapid multiplication of effort of well thought out constructs by removing the drudgery of lot of copy and paste coding and debugging and also synthesizing information across the disciplines as the natural world does not have the silos of the traditional academic disciplines. The folks who can see this future and adapt will be on the right side of the disruption and the others will be left behind.

Dawn of a golden era for computational science! Read More »

Happy Father’s Day 2025

Innovation, Materials, Science, Technology /

Happy Father’s Day! Especially to the scientists who are not getting much love these days and enduring a lot of uncertainty. I spent my time watching Cosmic Dawn by NASA (https://youtu.be/uSMGENDH_QI?si=1SVttqIqxP6oat50) with my kids – the fabulous story about James Webb Space Telescope (JWST). We don’t do enough in science to publicize successes.  JWST is a true testament to human intelligence and ingenuity, international cooperation across 14 countries, amazing perseverance and grit to pack the most complex telescope and sun shield into the cargo of a rocket and moving it into a orbit a million miles with a tolerance of few feet where it is unpacked remotely Origami style.  This is a story of 3 decades of relentless human pursuit starting from 1996. This project overcame 344 single point failures that were identified with only one chance. Just think about that probability and the rigor that went into something like this mission. It survived many black and gray swan events like Hurricane Harvey, COVID, many administrations and government changes across the world. Absolutely no room for error as it had to survive the brutal vibrations of the launch. Once deployed, the Origami of the unfolding of the telescope and sun shield a million miles away had to be perfect. The telescope itself is a human marvel but the more complex and underreported achievement is the sun shield with 1 mil membrane thickness that unfolds to tennis court size shield; five layers of those with precise separation to provide the thermal resistance to keep the telescope cold enough and protected from the sun. The data coming in is already remarkable that is changing our understanding of the universe: a) The universe is possibly older than 13.7 billion years  b) The data seems to contradict the Standard Model and current dark energy conceptualization  c) Getting us a glimpse of the formation of galaxies and getting us closer to the Big Bang  The benefits to humanity are much more than the insights into the origins of the universe and our own existence. The process led to 100s of innovations in materials, robotics and controls, most rigorous FMEA for extremely critical missions, and the most important human resource development. The new knowledge derived can create things in the future that we can’t imagine but the value of the output of the kids inspired is as boundless as the universe itself. We are reaping the benefits of the science and technology derived from the space explorers of the past pushing us in Star Wars parlance of going where no one has gone before – in this case through the eyes of JWST.

Happy Father’s Day 2025 Read More »