02 Mar, 2021
🎨 1. Showcase – A research team from ETH Zurich and Google introduced HiFiC short for High-Fidelity Generative Image Compression at NeurIPS last year. They use generative adversarial networks to create a state of the art lossy image compression system with astonishing results as seen below. An interactive demo is available online (hific.github.io).
📄 2. Paper – Elon Musk's Neuralink is working on a method to convert brain activity to text. This could help for example paralyzed people to communicate via writing. Their Brain-Computer-Interface (BCI) can produce 90 characters per minute!
Dr. Zsolnai-Fehér behind the YouTube channel Two Minute Papers dives into Neuralinks BCI paper "High-performance brain-to-text communication via imagined handwriting" (Paper on biorxiv.org, Two Minute Papers summary: 7 min on youtube.com).
💡 3. Use Case – Surveying the Skies with AI: Gravitational Lenses bend the path of light that travels towards, which affects the image that we can perceive from our earthly perspective. These gravitational lenses can be hard to discover, and here researchers have trained a deep ResNet to help them find these artifacts in space imagery (phys.org)
📰 4. Article – Srinivasa Ramanujan (1887 - 1920) was an Indian mathematician famous for making important contributions to the field without any previous formal education in mathematics (He was depicted in the book and 2015 movie "The Man Who Knew Infinity"). His notebooks were filled with theorems that turned out true, but were only proven decades later.
In comes AI: The Ramanujan machine wants to emulate its namesake by discovering new mathematical formula in an automated manner. As with the original Ramanujan, proving the discovered theorems will be left to its successors – (Read more at nature.com).
👁 5. Miscellaneous – Safari is the last major web browser (minus Firefox) to add Speech Recognition ability by default (cnet.com). Soon more websites will add custom voice commands to their online presences, following where Google has already started (9to5google.com).
💭 6. Tutorial – Quoc Le and Kip Kaehler from from Roblox describe how they sped up BERT inference by over 30x to serve 1+ billion daily requests with low latency (robloxtechblog.com).
💻 7. Code & Tutorial – Facebook AI's powerful wav2vec2 is state of the art speech-to-text transcription. It was recently added to the Huggingface Transformers library (huggingface.co).
💭 8. Article – Facebook improved its automated hate speech detection. Using their RIO and the self-supervised Linformer model (both published) they now detect 97% of hateful posts automatically before any human flags it (tech.fb.com).
💙 March 3 (online, German, 16:00) – State of AI – KI-Anwendungsideen in der Logistik – Unser Geschäftsführer Dr. Jürgen Stumpp präsentiert robuste KI Anwendungen für Supply Chain und lädt zur Diskussion mit Mitgliedern der BVL Bundesvereinigung Logistik. – Anmeldung bvl.de.
📅 March 3 (online) – Robert Bosch AI CON 2021 – 300 AI Experts from Industry and Academia get together. Welcome speakers from Carnegie Mellon University, Max Planck Institute and ETH Zurich. – Register here at bosch-ai.com.
17 Mar, 2021
📍 2021 AI Index released The Stanford Institute for Human-Centered Artificial Intelligence (HAI) released the newest edition of their annual report on AI impact and progress. Explore the full r…
04 Mar, 2021
Our current research project is about how goods get into the package - and into the "perfect package" at that. The aim is to use AI to help make the packing process more efficient and to optimize shi…