Links #3
Links #3
Orexin System: The Key for a Healthy Life
Sleep disorders, obesity, and aging: The role of orexin
Genetically encodable materials for non-invasive biological imaging. Protein-based contrast agents are neat. I wonder if there is a way to genetically modify neurons to produce these proteins and improve MRI imaging of the brain.
Coding metamaterials, digital metamaterials and programmable metamaterials
DNA-Templated Organic Synthesis: Nature's Strategy for Controlling Chemical Reactivity Applied to Synthetic Molecules. Cool concept, but I haven’t seen much follow-up work.
Bagging is an Optimal PAC Learner
Provably Faster Gradient Descent via Long Steps. It seems that cycling through different step sizes improves gradient descent. Surprising to see continued progress in convex optimization.
The Russian Roulette: An Unbiased Estimator of the Limit. The author also has a useful list of optimization inequalities and points to László Kozma’s inequalities cheat sheet.
Efficient and Near-Optimal Online Portfolio Selection. Seems like a performance improvement over other universal portfolio approaches. It’s funny that real hedge funds don’t seem to use these algorithms.
The end of this Casey Handmer post has a list of cool project ideas
The Interstellar Research Group posted many interesting talks:
X-ray and Y-ray Beam Interstellar Communication and Implications for SETI
Sunbeam: Near-Sun Statites as Beam Platforms for Beam Driven Rockets
Propelling Interstellar Exploration: Extreme-Aspect-Ratio Metamaterials in Post-Moore Era
One of the talks also pointed me to Jordin Kare’s Sailbeam concept. We’re quietly making progress on solar sails in general, and I’m excited for the era where we put these designs in space.
A ToughSF tweet on magnetotactic bacteria. This is a neat way to spatially organize artificial cells and drug delivery systems.
FPGA’s are pretty remarkable. A piece of hardware that can be programmed to do any computational task with performance close to ASIC’s lends itself to a lot of interesting applications. In addition to carrying out normal computations, they can be used for antennas, transmitters, physical random number generators, physical unclonable functions, reservoir computing, and have applications in hardware security and obfuscation. A selection of interesting papers:
FPGA-based Physical Unclonable Functions: A comprehensive overview of theory and architectures
FPGA Vendor Agnostic True Random Number Generator
HOP: Hardware makes Obfuscation Practical
An Overview of FPGA-inspired Obfuscation Techniques
Secure Public Verification of IP Marks in FPGA Design Through a Zero-Knowledge Protocol
Exploration of FPGA interconnect for the design of unconventional antennas
Towards a slime Mould-FPGA interface
EDIT: I came across more weird things you can do with FPGA’s including high-frequency trading, physical unclonable functions, and this excellent post listing more ideas.