AGI Hardware Paradigm Shifts and Neuromorphic Tech

chevron-icon
RFP Proposals
Top
chevron-icon
project-presentation-img
Expert Rating 4.0
Gabriel Axel Montes
Project Owner

AGI Hardware Paradigm Shifts and Neuromorphic Tech

Expert Rating

4.0

Overview

Integrated circuit designer and neuromorphic computing engineer Omomuyi Olajide and neuroscience-AI specialist and SingularityNET veteran Gabriel Axel Montes (Neural Axis) team up for a rigorous investigation of hardware paradigms and designs for advancing AGI. This proposal investigates neuromorphic processors, distributed systems, sparse computing, and others to optimize AGI workloads, focusing on Hyperon. By identifying viable and optimal configurations, the research aims to enhance efficiency, scalability, and cognitive synergy, addressing the needs of AGI and Hyperon components like PLN, MOSES, and ECAN while supporting decentralized compute systems for transformative AGI advancements.

RFP Guidelines

AGI related hardware

Complete & Awarded
  • Type SingularityNET RFP
  • Total RFP Funding $80,000 USD
  • Proposals 5
  • Awarded Projects 1
author-img
SingularityNET
Oct. 4, 2024

This RFP seeks to explore and evaluate innovative hardware paradigms, ranging from small-scale IoT devices to large, decentralized systems, to optimize AGI workloads in the OpenCog Hyperon framework and handle advanced processes such as cognitive synergy and hyperdimensional computing. The focus is on assessing hardware paradigms and emerging architectures (e.g. neuromorphic processors, associative processors such as TensTorrent’s APU, etc). The goal is to enhance computational efficiency, scalability, and cognitive synergy in AGI systems. Part of this should involve interacting with the Hyperon team who've built the existing and in-development MeTTa interpreters.

Proposal Description

Company Name (if applicable)

Neural Axis

Project details

Introduction


In this proposal, Omowuyi Olajide—expert integrated circuit (IC) design and neuromorphic engineer—and Gabriel Axel Montes—a neuroscientist and consciousness researcher, AI specialist, and SingularityNET veteran—, and Theo Valich—AI compute industry expert and SingularityNET partner ECOBLOX CEO are teaming up (see “About the Team” below) to conduct a rigorous examination of hardware paradigms for advancing AGI. Driven by experience and passion for the potential of hardware enhancements to advance cognitive capabilities, the team aims to make substantial contributions to advanced AI and AGI architectures and capabilities, including the Hyperon framework and its components. With Gabriel’s long-standing relationship with SingularityNET and the Hyperon team and Omowuyi and Theo’s expertise, we envision setting a strong foundation of recommendations for enhancing cutting-edge AI and AGI capabilities.


Research Plan

 

Achieving a scalable AGI system with the Hyperon framework requires a multifaceted hardware strategy. This includes leveraging massively parallel processing capabilities, deploying distributed computing infrastructures, integrating specialized hardware for neuro-symbolic processing, and adopting energy-efficient technologies. By aligning hardware configurations with the specific demands of AGI components, developing systems capable of performing at human-level intelligence and beyond becomes feasible.  The optimization of AGI systems like OpenCog Hyperon demands a holistic evaluation of cutting-edge hardware paradigms tailored to mixed workloads, cognitive synergy, and decentralized systems.

This research will commence with an in-depth literature review and comparison of hardware paradigms and solutions. These may include approaches such as neuromorphic and distributed systems processors, Application-Specific Integrated Circuits (ASICs) for Fully Homomorphic Encryption (FHE), sparse computing and heterogenous architectures, and others. Specific hardware/chips such as GSI and Tenstorrent APU, GPU’s, will be reviewed. This will be closely followed by an analysis of the relevance of hardware paradigms for AGI, including the performance of key AGI components. Among the potential found AGI-relevant analysis, the aim will be to evaluate the feasibility of the various hardware for AGI tasks, such as cognitive synergy, pattern matching, real-time inference, and others. These potential solutions and hardware capabilities will then be thoroughly examined for scalability, energy efficiency, reliability, and cost, as well as the potential to integrate the hardware solutions with AGI systems, including Hyperon

Based on the project proposal’s literature review and analysis of hardware solutions, we will identify use cases where AGI components, e.g. pattern matching or resource allocation, could benefit from certain hardware paradigms. Hyperon technologies will be evaluated against various hardware configurations for improved processing of the various cognitive modules. Emerging hardware technologies will be analyzed against existing solutions for the AGI components’ respective computational requirements. The project will allow recommendations to be made for integrating hardware into AGI systems, including the Hyperon stack, and how specialized hardware can support large-scale AGI computation. Consideration of rethinking codebases and compilers will be discussed with the Hyperon team. Furthermore, based on the distilled results of this research stage, recommendations for future research will be made for further experimentation with hardware solutions for AGI, e.g. for particular AGI research tasks.

Neuromorphic Computing and Hardware

Of special interest to the team are practical applications of hardware paradigms such as leveraging neuromorphic computing for:

  • Networking between GPUs (e.g. latency reduction, scalable interconnectivity);

  • Cybersecurity (e.g. anomaly detection, low-power IoT security);

  • Computer vision (e.g. efficient pattern recognition, energy-efficient edge processing)

  • Fully Homomorphic Encryption (FHE)-Optimized Hardware, for e.g., ensuring data privacy by allowing computations on encrypted data, critical for AGI systems operating in sensitive or decentralized environments;

  • Mixed signal In-memory compute accelerators, for integrating memory and computation at the hardware level for increased efficiency of, e.g. deep neural network inference;

  • Reconfigurable digital neuromorphic hardware, designed to emulate the brain’s neural networks, offering flexibility to adapt to various computational tasks.

These investigations will have utility for AI advancement more generally in addition to novel AGI-relevant solutions, and recommendations will be made for utilizing hardware configurations for various AI and AGI functions/tasks.

Resources


The requested project resources/budget reflect the time, expertise, and the value of the background knowledge of and relationships with the SingularityNET and Hyperon team that is anticipated to add significant value to the project proposal.

About the Team

 

Omowuyi Olajide

PhD(c); IC Design and Neuromorphic computing engineer
https://www.linkedin.com/in/omowuyi-olajide/ 

Omowuyi O. Olajide is a distinguished Integrated Circuit (IC) Design and Neuromorphic Engineer affiliated with the Department of Bioengineering and the Institute for Neural Computation at the University of California, San Diego. His expertise encompasses neuromorphic computing, Very-Large-Scale Integration (VLSI), quantum computing, mixed-signal circuits, computer engineering and programming, AI, and brain-machine interfaces. In a co-authored work, "Reconfigurable Event-Driven Spiking Neuromorphic Computing near High-Bandwidth Memory," presented at the IEEE Biomedical Circuits and Systems Conference (BioCAS), he explores the integration of spiking neuromorphic computing systems with high-bandwidth memory to enhance computational efficiency. He has led the development of an improved throughput for non-binary low-density parity-check decoders, as detailed in his first-authored article in Computer Engineering and Applications Journal. This research addresses advancements in error correction codes, which are vital for reliable data transmission in communication systems. Olajide's contributions to artificial intelligence extend beyond algorithm development to include the design of hardware that emulates neural networks for real-time, low-power decision-making. His work exemplifies a unique blend of creativity, technical expertise, and a relentless drive for innovation.

 

Gabriel Axel Montes


PhD; Neuroscientist, AI specialist
CEO, Founder - Neural Axis
https://www.linkedin.com/in/gabrielaxel/ 


Theo Valich

Advisor
CEO, Founder, ECOBLOX
https://www.linkedin.com/in/theovalich/ 

Theo Valich is a distinguished technology leader with extensive experience in high-performance computing (HPC) and data center solutions. As the Founder and CEO of Ecoblox, he leads the company's mission to develop sustainable, modular data centers that minimize environmental impact through innovative practices. An expert in both GPU and CPU technologies, Mr. Valich has been instrumental in the development of multiple national supercomputers. His career includes serving as CEO of a Swiss multifamily holding company, where he facilitated growth in emerging Middle Eastern markets. He has also held various leadership roles, including Chairman of the Board, CTO, COO, and executive advisor for companies across the U.S. and Europe.Mr. Valich began his career in virtual game development, contributing to Formula 1 racing titles. He later co-founded a digital media publishing company that focused on graphics technology, platforms, advanced video production, and central processing units. Under his leadership, Ecoblox has entered into a joint venture to develop a Carbon-Aware Routing Protocol, aiming to optimize and offset the carbon footprint of internet and data center operations. Mr. Valich continues to work closely with customers and supercomputing partners to advance Ecoblox's vision of being the greenest high-performance modular data center provider, fully offsetting environmental impact through sustainable practices and innovations.

 

Open Source Licensing

Custom

As this project is primarily a literature review, the literature review and Hyperon analyses will be made open-source.

Proposal Video

Not Avaliable Yet

Check back later during the Feedback & Selection period for the RFP that is proposal is applied to.

  • Total Milestones

    4

  • Total Budget

    $68,000 USD

  • Last Updated

    8 Dec 2024

Milestone 1 - Hardware paradigms & research plan

Description

This milestone represents commencement of the research setting the stage for further project work. It will include a research plan and early-stage preliminary findings about potential project directions around hardware approaches and their possible applicability to AGI.

Deliverables

1. Submission of a comprehensive research plan outlining the hardware paradigms and implementations to be explored. 2. Early (preliminary) identification and analyses of several potential hardware approaches—e.g. neuromorphic and distributed systems processors Application-Specific Integrated Circuits (ASICs) for Fully Homomorphic Encryption (FHE) sparse computing and heterogenous architectures etc.— and suggestions of their possible application areas within AGI systems. 3. Proposal of a detailed project plan including methodologies for evaluating the selected paradigms.

Budget

$17,000 USD

Success Criterion

A foundation for further project research will have been attained, where the team has gathered a scope of potential hardware approaches as well as possible AGI application areas, subject to refinement in later milestones.

Milestone 2 - Substantial research progress

Description

At this stage the research will have amassed adequate mapping of hardware paradigms and their capabilities. This will enable painting a picture of how to pragmatically leverage the hardware options to advance AGI. This stage precedes the identification of more specific use cases.

Deliverables

1. Further fleshing out of the research findings in a text document containing additional information and details on the identified quantum paradigms. 2. Additional theoretical and practical findings on how various hardware could improve AGI system performance integration feasibility scalability and/or energy efficiency. This stage may optionally begin looking at existing quantum hardware e.g. GSI and Tenstorrent APU GPU’s and/or others to note current market products that could be leveraged for AGI systems.

Budget

$17,000 USD

Success Criterion

Hardware literature review and AGI-relevant analysis have progressed substantially, evidenced by a lengthened research text document (with bibliographic references).

Milestone 3 - Continued deeper research progress

Description

This stage will advance project research to a deeper analysis of AGI-relevant hardware paradigms and their capabilities further identifying opportunities related to performance efficiency scalability reliability and so on. An evaluation of Hyperon technologies will have commenced with regard to how various hardware configurations could improve processing. Relevant hardware will be reviewed in particular products available on the market e.g. e.g. GSI and Tenstorrent APU GPU’s and/or others.

Deliverables

1. Updated documentation reflecting deeper analysis of hardware paradigms AGI-relevant analysis performance and other considerations; and how they may address shortcomings in existing state-of-the-art hardware. 2. Review of hardware products available on the market. 3. Partial identification of potential use cases for hardware paradigms in AGI including Hyperon/OpenCog and its components. 4. An evaluation of how hardware paradigms could improve processing in/of the Hyperon framework.

Budget

$17,000 USD

Success Criterion

At this stage, the research will have attained greater clarity over the applicability of the hardware paradigms to key AGI tasks, e.g. reasoning, pattern recognition, and resource allocation; and greater clarity over AGI performance; and opportunities for improving Hyperon processing capabilities.

Milestone 4 - Completion of Project Research

Description

This stage will have completed the scope of the “literature review and AGI-relevance analysis” phase of research setting the larger foundation for next-stage R&D using the insights from this research. This stage will provide clarity regarding opportunities presented by various hardware paradigms for specific AGI functions/tasks. This stage will have identified realistic use cases in AGI including Hyperon/OpenCog. Considerations on rethinking codebases and compilers to take advantage of hardware recommendations will be noted. Enablements from completion of this project for future research could include acquisition of relevant hardware to conduct experiments utilization of hardware paradigms and their functionalities to advance AGI functions/tasks including Hyperon components. If not covered in previous milestones this stage may optionally include analysis of any relevant combinations of examined hardware paradigms and quantum computing (QC) where applicable. This stage may include a high-level overview of possible cognitive enhancements for humans in the era of brain-computer interfaces (BCI) where humans and AGI become more closely coupled.

Deliverables

1. Detailed findings on the hardware paradigms’ and their functionalities’ impact on AGI performance scalability and computational efficiency. 2. Analysis on how particular hardware paradigms meet or may enhance the computational requirements of the various cognitive modules of AGI including Hyperon. 3. Identification of realistic use cases for particular hardware paradigms in advancing cutting-edge AI and AGI efforts including Hyperon/OpenCog. 4. Further analysis of results pointing out possible integrations or further study paths for the selected hardware paradigms and approaches. This may include potential suggestions for acquisition of hardware or cloud computing for experiments. 5. Noted relevant considerations on rethinking codebases and compilers to take advantage of hardware recommendations. 6. Optional: Analysis of any relevant combinations of examined hardware paradigms and quantum computing (QC) where applicable. 7. Optional: If not addressed in the previous milestone this stage would include a high-level explorative overview of how relevant hardware affordances may potentially enhance human (or human-AI–coupled) cognition that may open up in the future due to the ingress of brain-computer interfaces (BCI) and/or other cognitive horizons.

Budget

$17,000 USD

Success Criterion

This stage will have completed a one-to-one analysis of hardware paradigms and their functionalities matched with the functions/tasks of advanced AI and AGI; elucidated the state of the art of hardware paradigms; how they can optimally enhance AGI and Hyperon functions, tasks, and performance, efficiency, scalability, reliability, etc; offered relevant considerations on rethinking codebases and compilers to take advantage of hardware recommendations; identified realistic use cases for hardware paradigms in AGI, including within the Hyperon/OpenCog framework; (optionally) included an analysis of quantum capabilities that may be relevant to identified hardware paradigms, as applicable; and (optionally) intimated potential ramifications for human-AGI–coupled cognitive enhancements from hardware paradigms, such as neuromorphic computing, in the coming era of BCI.

Join the Discussion (0)

Expert Ratings

Reviews & Ratings

Group Expert Rating (Final)

Overall

4.0

  • Feasibility 3.7
  • Desirabilty 4.0
  • Usefulness 4.3

Proposal received high ratings from reviewers but experts ultimately selected another winner for strategic relevance. Strong potential interest in funding this work in a subsequent round.

  • Expert Review 1

    Overall

    4.0

    • Compliance with RFP requirements 4.0
    • Solution details and team expertise 4.0
    • Value for money 4.0
    Team suggests comprehensive hardware review with emphases on PRIMUS components. I would like to see more MeTTa compiler codebase research in order to benefit from certain chip architecture.

    Accept! Team suggests comprehensive hardware review with emphases on PRIMUS components. I would like to see more MeTTa compiler codebase research in order to benefit from certain chip architecture. Team expertise seems good and the best value for the money for given RFP.

  • Expert Review 2

    Overall

    4.0

    • Compliance with RFP requirements 5.0
    • Solution details and team expertise 5.0
    • Value for money 4.0
    It's a quality, solid proposal that meets the requirements of the RFP, and the team has the needed expertise

    It's an excellent proposal and the team is going to cast the net wide and also go deep. I have a slight concern that the team members are strong in cog sci / neurosci and in chip stuff, but none are really strong in technical AI... so the team will need close collaboration with the AI experts on the Hyperon team to really do this review in an appropriate way. OTOH Gabe is already collaborating on other projects so it's not a big leap to see this happening....

  • Expert Review 3

    Overall

    4.0

    • Compliance with RFP requirements 2.0
    • Solution details and team expertise 3.0
    • Value for money 5.0

    Solid evaluation framework and team. No discussion of Primus architecture-specific hardware though

feedback_icon