XMOHANEmbodied learning and original AI model architectures
Learning through interaction with the physical world, perception, memory, action and experience-driven adaptation.
A personal record of scientific research across distributed systems, embedded computing, artificial intelligence, and the physical world. Early research from 1999–2005 is presented alongside subsequent work in enterprise systems, deep learning, affective computing, AI hardware and applied research, and the current embodied-intelligence program at Robometrics® AGI. Each chapter is attributed to the researchers, organizations, and period in which the work originated.
At Robometrics® AGI, an independent research lab, my present research pursues embodied general intelligence for the physical world. The work investigates learning through experience, interaction between perception, memory and action, and adaptation in changing environments. These are research directions, not a claim of achieved AGI.
Learning through interaction with the physical world, perception, memory, action and experience-driven adaptation.
Research into bounded learning and methods of iterative improvement under the constraints of local computing.
Interpreting human emotional cues and studying proposed affect-like or digital-limbic mechanisms; subjective experience remains an open question.
The separate Robometrics® Machines programs—Amelia™ and aviation Copilot AI—explore constrained, advisory AI and human–machine interfaces.
Read the current research program, recent technical work and public research-related writings →
Beyond the early laboratories, my work has explored contextual video perception, natural-language search, deep learning, proposed AI chips, large-scale computing, human-centered AI, aviation and legal access. These programs span original research questions, experimental systems, applied engineering and research-related advisory work.
Video-based contextual perception, natural-language experience search and the vision of affect-aware robotic systems.
Aviation Copilot AI, Amelia™, human–machine interaction and computing under physical-world constraints.
Legal information systems, access to justice, accountability and responsible uses of AI in law.
AI architectures and infrastructure, semiconductors, cybersecurity, robotics and technology due diligence.
My first engineering position after graduating from IIT Roorkee was at HFCL R&D (1999–2000), programming microcontrollers in assembly language, sometimes without an operating system and with only a few kilobytes of RAM, while developing real-time cable-television network equipment. At Nortel Networks R&D in Gurugram (formerly Gurgaon), India (2000–2001), I developed and optimized enterprise-router software using a proprietary RTOS, as part of an internationally coordinated engineering effort. Some of the code I wrote ran on critical network infrastructure powering the New York Stock Exchange. I also researched mobile voice browsing. These experiences preceded my research at Brown University, HP Labs, MIT Media Lab, IBM Research Zurich and Intel Research Berkeley.
Explore the complete 1999–2005 research history, including HFCL and Nortel Networks →
Distributed data summaries, adaptive query routing and caching for decentralized information retrieval, with related technical publications.
Arrow Multicast, latency and packet-order experiments, and a peer-to-peer multiplayer gaming prototype.
Multipath routing, path quality, convergence, load balancing and sensor-network data aggregation.
Multi-channel software-defined AM reception, receiver architecture, signal processing and spectral displays.
Sensor-network experiments in multihop routing, energy efficiency and packet delivery.
An inside view of research instrumentation, experimental sensor boards and networked systems at Intel Research Berkeley.
Acoustic networking and hybrid communications for monitoring cargo containers in RF-constrained environments.
Bandwidth-adaptive multicast and resolution-aware information dissemination for pervasive monitoring.
Early voice browsing across mobile devices, proxy-based search, accessibility and WAP networks.
Cooperative traffic routing using vehicle-to-vehicle communication and local traffic information.
Further notes from the original research record, together with selected applied engineering and systems experience.
Cloud management, virtualization, technical partnerships, software licensing and enterprise systems integration.
Publications, reports and manuscripts →
About the researcher: scientific background, professional experience and creative practice →