Apparently, of the 195 Million the DoD allocated in University Research Funding Awards in 2022, more than half of them concerned AI or compute hardware research

post by mako yass (MakoYass) · 2023-07-07T01:20:20.079Z · LW · GW · 5 comments

This is a link post for https://www.defense.gov/News/Releases/Release/Article/2953234/department-of-defense-announces-university-research-funding-awards/

Contents

  Social Network-Transcendent Behavioral Dynamics Learning Dynamics and Detecting Causal Pathways in Coupled Online- Offline Systems
  Microelectronic Test Science Exploiting Latent Energy and Electromagnetic Radiation Comprehensive Minimally/non- invasive Multifaceted Assessment of Nano-/ microelectronic Devices (CoMMAND)
  Understanding Neural Systems Integration for Competent Autonomy in Decision and Control FLAP: Fast, Lexicographic Agile Integrates Decision and Control in a Spike-Resolved, Sensorimotor Program
  Bio-architected Responsive Materials with 3D Nanoscale Order 3D Nanomaterial Architectures with Programmable Reconfiguration States and Functions
  ELECTROBIOLOGY: Electronic Control of Biological Communication Faster, More Efficient, and Hybrid Computation in Microbial Bioelectronic Systems ARO Rice University
  Novel Routes to Majorana Qubits for Topologically- Protected Quantum Information Epitaxial phase-biased Josephson junctions
  Molecular Doping of Organic Electronic Materials Next Generation Molecular Dopants for Organic Electronics: From Fundamentals to New Device Concepts
  Self-learning for Real-world Perception Intentional multi-modal self-learning to perceive and understand the real world
None
5 comments

Suggests interesting things about what the Department of Defense is focusing on, so I thought I should mention it.

The awards concerned "pursuing basic research spanning multiple scientific disciplines", which makes it pretty interesting that three fifths of them are either directly AI and compute hardware research, or have applications to compute hardware research.

From a merit-based review of approximately 340 proposals received, a panel of experts narrowed the proposals to a subset, from which the 28 final awards were selected

Take a look at the table of project topics. Bold means probably directly AI/compute, italic means probably primarily applicable to AI/compute.

Social Network-Transcendent Behavioral Dynamics Learning Dynamics and Detecting Causal Pathways in Coupled Online- Offline Systems

Microelectronic Test Science Exploiting Latent Energy and Electromagnetic Radiation Comprehensive Minimally/non- invasive Multifaceted Assessment of Nano-/ microelectronic Devices (CoMMAND)

Cavity Molecular Polaritons MURI: Polariton Chemistry

Effects of Radiation Damage on Performance of Wide-Bandgap Electronics REDESIGN: Radiation Effects preDiction through Experimentally validated Simulation of Gallium Nitride

Understanding Neural Systems Integration for Competent Autonomy in Decision and Control FLAP: Fast, Lexicographic Agile Integrates Decision and Control in a Spike-Resolved, Sensorimotor Program

Nonlinear Optical Material Design with Extreme Interband Nonlinearities Atomic-scale Manipulations of Interband Optical Nonlinearities - (AtoMs➝IONs)

Synthetic Quantum Matter Reimagining Atoms and Photons in SYnthetic, DYnamical, and INteracting Quantum matter

Composability of Synthetic Biological Circuits Rules of Composition in Synthetic Biology Across Scales of Complexity: Theory and Tools

Bio-architected Responsive Materials with 3D Nanoscale Order 3D Nanomaterial Architectures with Programmable Reconfiguration States and Functions

Topological Seeds of Complex Response in Materials Programming Multistable Origami and Kirigami Structures via Topological Design

Connectivity and Transport in Disordered Hyperuniform Networks Transport in Disordered Hyperuniform Systems and Networks

Uncovering the Underlying Neurobiological Mechanisms of Cognitive Fatigue Understanding and Predicting Cognitive Fatigue across Multiple Timescales, Distinct Aspects of Cognition, and Different Individuals with Multiscale Whole Cortex Models

Gut-Neuronal Signaling Through Polymeric Mucin via Chemical Probes and Imaging Mucin-mimetic Interventions to Modulate the Gut-Brain Axis

ELECTROBIOLOGY: Electronic Control of Biological Communication Faster, More Efficient, and Hybrid Computation in Microbial Bioelectronic Systems ARO Rice University

Novel Routes to Majorana Qubits for Topologically- Protected Quantum Information Epitaxial phase-biased Josephson junctions

Molecular Doping of Organic Electronic Materials Next Generation Molecular Dopants for Organic Electronics: From Fundamentals to New Device Concepts

Learning from Hearing Neurobehavioral, Physiological, and Computational Processes of Auditory Object Learning in Mammals

Hydrodynamics of Fish Schooling Revealing the Hydrodynamic Principles of Three-dimensional Fish Schools; From Biology to Schooling Robotics

Self-learning for Real-world Perception Intentional multi-modal self-learning to perceive and understand the real world

Fundamental Non-equilibrium Processes in Weakly Ionized Hypersonic Flows Development of Validated Hypersonic Plasma Kinetics Models Including Atomic Excitation

Initio Understanding of Detonation Based Combustion in Multiphase Mixtures Multiphase Detonation of Liquid Aeropropulsion Fuels

Bioinspired Design of Energy- Self Sufficient Multi-functional Soft Material Systems SERPENT: Self EneRgetic Propulsion ENTity

Systems-Level Foundations for Agile, Dynamic, and Ad Hoc Human Autonomy Teams HUDDLE: Human Autonomy Teaming in Uncertain and Dynamic Environments

Environmental DNA-based Monitoring of the Marine Environment (ED- MON) MMARINeDNA: Marine MAmmal Remote detection via INnovative environmental DNA sampling ONR University of Washington

Flat Bands Beyond van der Waals Materials Topological Flat Bands for Correlated Electron Systems

There are multiple entries in computing with biological substrates.
The materials science stuff, "Polariton Chemistry", "Reimagining Atoms and Photons in SYnthetic, DYnamical, and INteracting Quantum matter", etc, this does all seem to be gesturing towards potential applications in compute hardware, and that's where I'd expect any potential applications to concentrate in practice?

I wonder how these projects have progressed over the past year.

5 comments

Comments sorted by top scores.

comment by trevor (TrevorWiesinger) · 2023-07-07T08:03:50.181Z · LW(p) · GW(p)

I'm surprised that this has so few upvotes 7 hours after being posted, the potential implications are big and it is clearly worth further research. There's still the possibility that DoD funding is less indicative of US technology policy/priorities than government R&D spending as a whole, but I wouldn't put too much stock in that (e.g. the DoD money could be the 20% of the money that does 80% of the results, or at least 80% of indication of US technology priorities/steering).

Have you considered posting this on EAforum? There seems to be more interest in policy there, even if the average EAforum user is less reliable than the average lesswrong user.

Replies from: MakoYass
comment by mako yass (MakoYass) · 2023-07-07T16:34:37.208Z · LW(p) · GW(p)

I'm surprised that this has so few upvotes 7 hours after being posted

Could be because I posted it in California's late evening. Btw, lightcone, if you still haven't addressed biased scoring due to the timing of posts, that would be really pathetic, and you need to.
Relatedly, I urge you to trial sorting new posts according to prediction markets over their eventual score ratios instead of continuing with a system where early scores (and so, also, eventual scores) are basically random.

Have you considered posting this on EAforum?

I'd want to see it investigated a bit more, first. I'm not sure whether this reflects a broader prioritization of compute in DoD or if this awards process was just in a quirky mood, or if academic research already has this focus, there are probably more thorough investigations into DoD research interests out there.
I also don't have a good intuition for whether this kind of blue sky research leads anywhere, this could just be representative of the least efficient part of the DoD or something and we might still expect most progress to come out of industry.
I also don't think there are AI people who hang out in the EA forum and not lesswrong.

Replies from: ViktorThink
comment by MrThink (ViktorThink) · 2023-07-07T21:52:49.800Z · LW(p) · GW(p)

I think one reason for the low number of upvotes was that it was not clear to me until the second time I briefly checked this article why it mattered.

I did not know what DoD was short for (U.S. Department of Defense), and why I should care about what they were funding.

Cause overall I do think it is interesting information.

Replies from: winstonBosan
comment by winstonBosan · 2023-07-07T23:24:42.704Z · LW(p) · GW(p)

Viktor has a point here - the title is informative, but not well optimized (perhaps intentionally) for attracting eyeballs. 

Something akin to:
Military and AI Compute: DoD's 100 million cheque and what did it get for them?  

Might do the trick a bit better.

Replies from: MakoYass
comment by mako yass (MakoYass) · 2023-07-07T23:28:11.634Z · LW(p) · GW(p)

"perhaps intentionally" I was going to concede no it wasn't intentional, but your suggested title complicated that a bit, I definitely would not want to use that title, xD we need to be careful to avoid making arms-racey declarations.