211
WebGPT: Improving the factual accuracy of language models through web browsing
We’ve fine-tuned GPT-3 to more accurately answer open-ended questions using a text-based web browser.
New
The latest blog posts to take off on Hacker News. 301 posts.
212
213
214
The Scaling Hypothesis (2021)
On GPT-3: meta-learning, scaling, implications, and deep theory. The scaling hypothesis: neural nets absorb data & compute, generalizing and becoming more Bayesian as problems get harder, manifesting new abilities even at trivial-by-global-standards-scale. The deep learning revolution has begun as foretold.
215
Removing blur from images – deconvolution and using simple image filters
Left: Original image. Center: Blurred through convolution. Right: Deconvolution of the blurred image. In this post, we’ll have a look at the idea of removing blur from images, videos, or games thro…
216
I made an AI write a story about AI then I made another AI illustrate it
I was walking home from school by myself when I saw it. A robot, just standing in the middle of the street.
217
“Amateur” programmer fought cancer with 50 Nvidia Geforce 1080Ti
Not All Heroes Wear Capes
218
219
220
My deepfake DALL-E 2 vacation photos passed the Turing Test
I went on a real vacation, and created photos of a fake vacation using DALL-E 2. Most people couldn't tell the difference.
221
Accelerated PyTorch Training on M1 Mac
In collaboration with the Metal engineering team at Apple, we are excited to announce support for GPU-accelerated PyTorch training on Mac. Until now, PyTorch training on Mac only leveraged the...
222
NaturalSpeech: End-to-end text to speech synthesis with human-level quality
Text to speech (TTS) has made rapid progress in both academia and industry in recent years. Some questions naturally arise that whether a TTS system can achieve human-level quality, how to define/judge that quality and how to achieve it. In this paper, we answer these questions by first defining the human-level…
223
AI and Machine Learning – The Basics
Artificial Intelligence is a once-in-a lifetime commercial and defense game changer (download a PDF of this article here) Hundreds of billions in public and private capital is being invested in Art…
224
We asked 100 humans to draw the DALL·E prompts
Where do human artists fit in a world of rich, creative AI? We asked 100 Surgers to draw the DALL-E prompts.
225
226
DeepMind: A Generalist Agent
Explore a selection of our recent research on some of the most complex and interesting challenges in AI.
227
BERTopic: The Future of Topic Modeling
BERTopic takes advantage of the superior language capabilities of (not yet sentient) transformer models and uses some other ML magic like UMAP and HDBSCAN to produce what is one of the most advanced techniques in language topic modeling today.
228
Cloud TPU VMs are generally available
Cloud TPU VMs with Ranking & Recommendation acceleration are generally available on Google Cloud. Customers will have direct access to TPU host machines.
229
Recognize Digits Using ML in Elixir
Machine learning allows you to solve once unimaginable problem. Elixir's support for machine learning through Nx and Axon let us do impressive thing while staying in our favorite language.
230
Jax vs. Julia (Vs PyTorch)
A while ago there was an interesting thread on the Julia Discourse about the “state of machine learning in Julia”. I posted a response discussing the differences between Julia and Python (both JAX and PyTorch), and it seemed to be really well received! Since then this topic seems to keep coming up, so I thought I’d…
231
Learning to think critically about machine learning
Graduate students are helping to infuse ethical computing content into MIT’s largest machine learning course, as part of the Social and Ethical Responsibilities of Computing initiative.
232
233
I built an open-source, AI-powered solar panel that's 95% optimal
This research demonstrates a model-free approach to optimize the energy produced by a dual-axis solar panel using reinforcement learning. Specifically, a softmax actor-critic agent optimizes energy production in a simulated, dynamic lighting environment which is generated from real power data. All s
234
Tackling multiple tasks with a single visual language model
We introduce Flamingo, a single visual language model (VLM) that sets a new state of the art in few-shot learning on a wide range of open-ended multimodal tasks.
235
An automatic theorem proving project
I am very happy to say that I have recently received a generous grant from the Astera Institute to set up a small group to work on automatic theorem proving, in the first instance for about three y…
236
The Machine Learning Job Market
I was on the job market recently, open to suggestions on what I should do next. I'm deeply thankful to all those who reached out and pitched their companies, and also to people who shared their wisdom of how they wanted to navigate their own careers.
237
Amazon releases 51-language dataset for language understanding
MASSIVE dataset and Massively Multilingual NLU (MMNLU-22) competition and workshop will help researchers scale natural-language-understanding technology to every language on Earth.
238
How DALL-E 2 Works
How does OpenAI's groundbreaking DALL-E 2 model actually work? Check out this detailed guide to learn the ins and outs of DALL-E 2.
239
Doing small network scientific machine learning in Julia faster than PyTorch
Doing small network scientific machine learning in Julia 5x faster than PyTorch ...
240
K-Nearest Neighbors
One Machine Learning algorithm that relies on the concepts of proximity and similarity is K-Nearest Neighbor (KNN). KNN is a supervised learning algorithm capable of performing both classification and regression tasks.