@Fra_Pochetti I did a quick comparison between deploying PyTorch via Lambda Layer vs via Docker and found a slight improvement in cold start times for the Docker version
@jeremyphoward@bradfitz something like Cassandra gives you tunable consistency with its replication factor config. It has fast writes, but reads will be slower than an in-memory store
@totteh I've thought of something similar. An interface like https://t.co/FhFMVDKExx to crowd-source layman's terms explanations of sections of papers and link them to corresponding code in official implementations. Author approved/written comments would be displayed prominently.
@nitishmalhotra9@rustlang@golang The Option enum is Rust’s way of handling values that could potentially be empty (or null/nil in other languages). By using Option the compiler is able to force you to handle the case in which a value is empty/null rather than getting those null pointer exceptions at runtime
Every year, we partner with a dozen startups and work with them heads-down over the summer.
Today, we're announcing our 2020 Accelerator Program. Learn more and apply here: https://t.co/BZyFEmG2jm
How lighting conditions can induce a different per-frame exposure time in a camera and cause a stroboscopic + rolling shutter effect, well visualized by this gif [source: https://t.co/s0yYOfIAnF]
A nice fact I like: Every matrix corresponds to a graph, and so familiar things (e.g. matrix multiplication) have nice pictures! Another nice fact: joint probability distributions *also* correspond to graphs. They have telling pictures, too. New blog post! https://t.co/EP67IhGAr0
For the last few years, one of the staples of my research has been visualizing individual neurons in vision models. But that's only a partial picture -- neurons work together.
Activation Atlases are a way to explore the space neurons jointly represent.
https://t.co/TEa9WuYBMg
Applying reinforcement learning to environments with sparse and underspecified rewards is an ongoing challenge, requiring generalization from limited feedback. See how we address this with a novel method that provides more refined feedback to the agent. https://t.co/PdjltHbgbH
It amazes me when people think numbers & math are somehow impervious to inaccuracy, bias, manipulation. Algorithmic bias is very real & I talk to AI researchers *weekly* about their efforts to solve this problem. Here's also my story about this topic. https://t.co/ud6X5B4tQc