AI, MLOps, and Robotics #18

This Week in AI, MLOps, and Robotics: Hacking LLMs at DEF CON, Geoffrey Hinton discusses dangers of AI, The White House meeting for responsible AI, YOLO-NAS Achieves State-of-the-Art Performance, a State-of-the-Art LLM for code, Why teaching robots to blink is hard but important, sentence embeddings can leak more expected, and much more!

My next Event (May 10th):

WhyLabs and Databricks: Simplifying ML Monitoring in Complex Environments

Register on Eventbrite ↗

?? News & Stories:

• DEF CON to set thousands of hackers loose on LLMs - Read on the Register ↗
• Geoffrey Hinton discusses dangers the developing technologies pose to society - Read/Watch on PBS ↗
• How robotics can improve the environment - Read on Robotics & Automation ↗
• Readout of White House Meeting with CEOs on Advancing Responsible Artificial Intelligence Innovation - Read on Whitehouse.gov ↗
• Artificial Intelligence comes with risks. How can companies develop AI responsibly? - Read on NPR ↗
• Can Deep Learning Improve My Manufacturing Process? - Read on Quality Magazine ↗
• Why teaching robots to blink is hard but important - Read on BBC ↗
• AV fever has cooled off, but driverless cars aren’t going away - Read on the Verge ↗

??? Developer Tools & Resources:

• YOLO-NAS Achieves State-of-the-Art Performance on Object Detection Using Neural Architecture Search - Read on the Deci AI blog ↗
• IndoorSim-to-OutdoorReal: Learning to navigate outdoors without any outdoor experience - Read on the Google AI blog ↗
• Segment Anything – A Foundation Model for Image Segmentation - Read on LearnOpenCV ↗
• StarCoder: A State-of-the-Art LLM for Code - Read on HuggingFace ↗
• Latest NVIDIA Graphics Research Advances Generative AI’s Next Frontier - Read on the NVIDIA blog ↗
• Scaling deep retrieval with TensorFlow Recommenders and Vertex AI Matching Engine - Read on the TensorFlow blog ↗
• Accelerated Image Segmentation using PyTorch - Read on the PyTorch Blog ↗
• BigCode project: Code-generating LLMs boosted by Toloka's crowd - Read on the Toloka blog ↗
• MaMMUT: A simple vision-encoder text-decoder architecture for multimodal tasks - Read on the Google AI blog ↗
• Training the YOLOv8 Object Detector for OAK-D - Read on PyImageSearch ↗

??? Upcoming Events (Ordered by date):

• From Data to Deployment |?Online, May 9, 3:00am PT?-?Register on Eventbrite ↗
• Workshop: LLM Ops and AI Agents?| Online, May 9, 9:00am PT -?Register on AI Camp ↗
• Simplifying ML Monitoring in Complex Environments?| Online, May 10, 10:00am PT?-?Register on Eventbrite ↗
• Build and Monitor Computer Vision Models | Online, May 17, 10:00am PT - Register on Eventbrite ↗
• ML Meetup: LLMs with LangChain?| Austin, May 17, 5:00pm PT -?Register on AI Camp ↗
• Intro to ML Monitoring: Data Drift, Quality, Bias and Explainability?| Online, May 24, 3:00pm PT-?Register on Eventbrite ↗

??? AI Podcasts & Interviews:

• Large models on CPUs | Practical AI - Listen on Changelog ↗
• Hyperparameter Optimization through Neural Network Partitioning with Christos Louizos - Listen on TWiML ↗
• Pandas for Data Analysis and Visualization - Listen on SuperDataScience ↗
• Lukas Biewald of Weights and Biases: solving pain-points in AI - Listen on the Robot Brains ↗
• Cost/Performance Optimization with LLMs [Panel] - Listen on MLOps Community ↗

?? Recent ML Research Papers:

• Tracking through Containers and Occluders in the Wild - Read on arXix ↗
• NeuralEditor: Editing Neural Radiance Fields via Manipulating Point Clouds - Read on arXix ↗
• Personalize Segment Anything Model with One Shot - Read on arXix ↗
• Improving Code Example Recommendations on Informal Documentation Using BERT and Query-Aware LSH: A Comparative Study - Read on arXix ↗
• Sentence Embedding Leaks More Information than You Expect: Generative Embedding Inversion Attack to Recover the Whole Sentence - Read on arXix ↗

Hello, world! Every week I keep track of the most interesting news, resources, events, podcasts, and papers I find around AI, MLOps, and robotics. Sometimes I sprinkle in my opinions. Let me know what I missed this week in the comments!

You Can Support This Newsletter By:

?? Leaving a comment with anything you found interesting this week!

??? Joining the discussion in the?Robust & Responsible AI Slack!

? Giving the?open-source ML monitoring project, I work on a?GitHub star!

Connect with or follow me (Sage Elliott) for more content on machine learning, MLOps, and robotics. Have a great week! ??