Top all-time · #16061
Frank’s Ramblings
Personal blog · frankzliu.com/blog · on HN as mrintellectualaidata
Frank’s Ramblings is a personal blog at frankzliu.com about ai and data. Since 2022, 4 different Hacker News readers have linked it 15 times, most often “Understanding Neural Network Embeddings”. The last post was in 2023, but the archive is still widely read.
Too few sampled comments (1) to say how often HN recommends it.?
Why HN links it
Only one comment appears, an Ask HN answer describing an image recognition pipeline (bounding boxes, embeddings, similarity search). It does not mention this blog or its writing.
Summarised by Claude from a sample of 1 HN comments that link here. Each comment is marked recommending, critical or neutral (citing without judging); how this works.
Professional presser of buttons on computer keyboards.
Most-linked posts?
- Understanding Neural Network Embeddings2 people · 2 mentions · first linked 2022-05-04
- My Experience Working and Living in China2 people · 2 mentions · first linked 2022-02-26
What Hacker News says?
This depends mostly on what kind of search engine you're trying to build. I unfortunately won't be able to point you towards courses, but there are tons of great resources online to help you get started. Search engines are a fairly broad topic, and a lot of it depends on the _type of data_ that you want to build a search engine for. If you're looking towards more traditional, Google/Yahoo-like search, Elasticsearch'…
Comment linking /blog/a-gentle-introduction-to-vector-databases
on “Ask HN: I want to dive into how to make search engines”Submitted A16Z blogs are just glorified marketing as a story · 278 points
HN discussionAgreed. ViTs are better if you're looking to go multimodal or use attention-specific mechanisms such as cross-attention. If not, there's evidence out there that ViTs are not better than convnets for small networks and at scale ( https://frankzliu.com/blog/vision-transformers-are-overrated ).
Comment linking /blog/vision-transformers-are-overrated
on “Convolutional Neural Networks for Visual Recognition”Submitted A gentle introduction to vector databases as a story · 133 points
HN discussionUsually this is done in three steps. The first step is using a neural network to create a bounding box around the object, then generating vector embeddings of the object, and then using similarity search on vector embeddings. The first step is accomplished by training a detection model to generate the bounding box around your object, this can usually be done by finetuning an already trained detection model. For this…
Comment linking /blog/understanding-neural-network-embeddings
on “Ask HN: Any good self-hosted image recognition software?”Submitted My Experience Working and Living in China as a story · 93 points
HN discussionSubmitted Vision Transformers Are Overrated as a story · 67 points
HN discussionSubmitted My Experience Working and Living in China as a story · 22 points
HN discussionSubmitted Pandemic Stories from China as a story · 8 points
HN discussionSubmitted Evading Bureaucracy: Making a PDF Look Hand-Signed and Scanned as a story · 8 points
HN discussionSubmitted Understanding Neural Network Embeddings as a story · 3 points
HN discussionSubmitted Vision Transformers Are Overrated as a story · 2 points
HN discussionSubmitted Vision Transformers are Overrated (2023) as a story · 2 points
HN discussionSubmitted Vision Transformers are Overrated (2023) as a story · 2 points
HN discussionSubmitted Making Machine Learning More Accessible for Application Developers as a story · 1 points
HN discussion