Amelia's Weekly Fish Fry

Adventures in Machine Learning: From Logic Simulation to Exoplanet Identification

Machine learning takes center stage in this week’s Fish Fry podcast! First, we examine a new machine learning algorithm developed by researchers at the University of Warwick Department of Physics and the Alan Turing Institute. We take a closer look at the datasets used in this new algorithm and why this groundbreaking research will help confirm the existence of exoplanets and significantly speed up our understanding of the universe. Keeping with our machine learning theme, Paul Cunningham (Cadence Design Systems) also joins us this week to discuss the details of Cadence’s new machine learning-optimized Xcelium logic simulation and how machine learning can accelerate regression throughput.

 

The ElephantEdge Challenge: Developing The World's Most Advanced Elephant Tracking Collar

This week’s podcast is all about The ElephantEdge Challenge and how you can help build the world’s most advanced elephant tracking collar. Adam Benzion (Co-Founder - Hackster.io) joins us to discuss the details of The ElephantEdge Design Challenge, what The OpenCollar initiative is all about, and how you can get involved to help save this vulnerable species from extinction.

Running the Long Game: New RISC-V Based SoCs and Using Technosignatures to Find Alien Life

It’s a bird, it's a plane, it's a RISC-V Fish Fry! In this week’s podcast, we first investigate how the search for alien life just got a lot more interesting with the help of a unique grant from NASA, the SETI institute, and a team of esteemed researchers from across the United States. We take a closer look at the details of this new research project and how it is hoping to find signs of alien life on distant exoplanets using technosignatures. Also this week, Sammy Cheung (CEO - Efinix) joins us to discuss how Efinix is blending the worlds of RISC-V and reconfigurable computing and the details of  their new family of software-defined SoCs based on the RISC-V core.

 

Hearing is Believing: Artificial Intelligence, Cough Patterns, and Stemming the Tide of COVID-19

“One person's data is another person's noise.” - K.C. Cole

This week’s Fish Fry podcast is all about noisy neurons, cough signature identification, and how AI can help us stem the tide of COVID-19. First up, we check out new research from the German Center for Neurodegenerative Diseases that may have finally unlocked the mysteries of memory retention in people afflicted with Alzheimer’s disease with a little help from a bunch of mice and some very noisy neurons. Also this week, Chris Rogers (CEO - SensiML) joins us to chat about a new health monitoring solution to help fight the COVID-19 pandemic. We discuss the role of artificial intelligence in this solution and how this new platform may help detect symptoms earlier than before with identification of cough sounds.

 

Building an AI Ecosystem: You Can’t Do it All By Yourself

We’re taking on industrial automation and AI at the edge in this week’s podcast! First, we take a closer look at a new deep-learning framework developed at UC Santa Cruz that can identify and classify galaxies and stars by analyzing astronomical image data pixel by pixel. Steve Cammish (VP of Edge Solutions, ADLINK) also joins us this week to discuss the biggest challenges of the artificial intelligence in the industrial arena and the details of the Vizi-AI™ Development Starter Kit for Industrial Machine Vision AI at the Edge.

Calling All Innovators: Why You Should Join the Detect and Protect Challenge

In this week’s podcast, we take a closer look at The COVID-19 Detect & Protect Challenge. Hackster.io co-founder Adam Benzion and I chat about why The United Nations Development Programme and a multinational group of companies got together to create this challenge, the goals, prizes, and details associated with this design challenge, and most importantly, how your open source designs can help make the world a better place. Also this week, we check out new flexible sensor technology developed at MIT that is hoping to make vital sign monitoring a whole lot easier. 

 

Systems and Bodies on Chip

In this week’s Fish Fry podcast, we take on systems on chips, body on chips, and much more!  First up, we check out how Havard’s Wyss University is hoping to change the future of modern drug development and approval. We take a closer look at how this team has pieced together ten “Organ Chips” to create a fully functioning body-on-chip platform and how this new BoC system can give us comprehensive new insights into how prospective drugs will behave throughout the human body. Also this week, I chat with Yorgos Koutsoyannopoulos (Ansys) about the challenges of chip design for 5G, radio-frequency integrated circuit design workflows and how we can all avoid electromagnetic crosstalk nightmares. 

 

Inventing Actionable Intelligence

Actionable intelligence and inference at the edge takes center stage in this week’s Fish Fry podcast. First up, we take a closer look at how drones can be taught to echolocate (like bats and dolphins) with a little help from a speaker, four microphones and a whole lot of math. Next, Nigel Forrester (Concurrent Technologies) and I chat about radio frequency signal intelligence, the benefits of a heterogeneous resource pool, and the details of Concurrent Tech's new artificial intelligence accelerator board.

 

Algorithmic Acceleration and The Next Phase of Moore’s Law

In this week’s Fish Fry podcast, we take a closer look at deep learning model acceleration and the future of Moore’s Law. First up, Tom Smelker (Mercury Systems) and I discuss modular chip design and why he thinks it will be crucial to the next phase of Moore’s Law. Next, we investigate AI hardware and algorithm acceleration with Dr. Mohamed Bergach and Marc Littlefield (Kontron). To finish things up, we take a closer look at the unique design challenges of micro air vehicles and how new research from China's Changzhou and Jiangsu universities could make a sun-powered flapping wing a reality.

New Kid on the Block

This week’s podcast is one part frog embryo, one part embedded design, and one part crystal ball! To start things off, we take a closer look at how a team of research scientists have created a new form of life (a living, programmable organism called a Xenobot) with a little help from African frog embryos, a algorithm that simulates designs for new life-forms, and the Deep Green supercomputer cluster at the Vermont Advanced Computing Core. Also this week, Dunstan Power (ByteSnap Design) and I discuss a variety of design trends for 2020, including the rise of RISC-V, the role of security in embedded design, and why he expects to see a big push into machine learning-driven by tools from Xilinx and Intel.