machinelearning

5 ms·

1.
Garbage In, Garbage Out: The Case for Better Robot Data Understanding(huggingface.co)
1 points by machinelearning 10mo ago|0 comments
2.
by machinelearning 10mo ago
Training is expensive so I wouldn't necessarily call it "pragmatic". - the tool's goal is actually to provide a lightweight, practical way to avoid wasting training cycles on bad data. Evals for robotics are also expensi
3.
Show HN: Automate robot data quality improvement(github.com)
9 points by machinelearning 11mo ago|2 comments
4.
A toolkit for improving the quality of your LeRobot datasets(github.com)
1 points by machinelearning 11mo ago|0 comments
5.
Show HN: LeRobot Episode Scoring Tool(github.com)
1 points by machinelearning 1y ago|0 comments
6.
by machinelearning 1y ago
"Most ways of constructing accurate models mathematically don’t actually work" > This is true for almost anything at the limit, we are already able to model spatiotemporal dynamics to some useful degree (see: progress in VLAs,
7.
by machinelearning 1y ago
> The reason is that it makes the essay easier to read. It's less work to read writing that flows well. How does that help the writer? Because the writer is the first reader Note that the writer's judgement only serves as an in
8.
by machinelearning 1y ago
So one thing to note is that the essay mentions that it refers specifically to "writing that is used to develop ideas" vs. "writing meant to describe others ideas". The way I interpret this is that it refers to claims th
9.
by machinelearning 1y ago
The issue with this article is that it is very imprecise. Are the standards for whether something “sounds bad” based on the average person’s reading or the intended audience. In its most general form (how the median article sounds to the me
10.
by machinelearning 1y ago
> “So here we have the most exciting kind of idea: one that seems both preposterous and true.” Am I missing something or is the “seems true” part taking too many liberties here? If anything, as described in the previous few sentences, th
11.
by machinelearning 1y ago
What's a use case are you interested in re: video?
12.
by machinelearning 2y ago
This is a good problem to solve but the approach is wrong imo. It has to be done in a hierarchical way to know what you attended to + full context. If the differential vector is being computed with the same input as the attention vector how
13.
by machinelearning 2y ago
No, you've only discussed the Retrieval part of RAG, not the generation part. The current workflow is to use the embedding to retrieve documents then dump the text corresponding to the embedding into the LLM context for generation. Oft
14.
by machinelearning 2y ago
Nope that’s not how most rag systems work today. I looked at pgvector and couldn’t find anything similar. Do you have a link? Or maybe you misunderstood what I was taking about
15.
by machinelearning 2y ago
Both RAG and infinite contexts in their current states are hacks. Both waste compute because you have to re-encode things as text each time and RAG needs a lot of heuristics + a separate embedding model. Instead, it makes a lot more sense t
16.
by machinelearning 2y ago
This is a simple version of the tree search approach that people suspect Q* is
17.
by machinelearning 3y ago
As someone who studied math, I approve of this video. Summary available here: https://www.askyoutube.ai/share/6533668f27f67cda427ab90f
18.
by machinelearning 3y ago
https://www.askyoutube.ai/share/6521aa6077733b7c0ad24b55
19.
by machinelearning 3y ago
Its much faster now and a share function has been added for pre-loaded examples
20.
by machinelearning 3y ago
Added
21.
by machinelearning 3y ago
Added
22.
by machinelearning 3y ago
That's reasonable, planning to add it tomorrow
23.
by machinelearning 3y ago
Thanks for the feedback, I haven't done much optimization in regards to the summary. The screen real-estate of the summary has to be small so it is limited in the amount of information it can display but your example could be a good st
24.
by machinelearning 3y ago
Glad that it worked well for you! I don't have any general advice that would be useful but that sounds like an interesting project
25.
by machinelearning 3y ago
Fair point, will add this tomorrow
26.
Show HN: Askyoutube – Ask YouTube anything(askyoutube.ai)
18 points by machinelearning 3y ago|15 comments
27.
by machinelearning 3y ago
In other words, a ‘hack’
28.
by machinelearning 3y ago
If you're trying to use this and don't get the sign up email, check your spam folder. Gmail seems to auto-categorize the email as spam
29.
by machinelearning 4y ago
The tech that you know has changed. We've moved well beyond programming languages and frameworks. Once you adapt to the frontier, your eyes will be opened to the revolutionary technology being developed right under your nose.
30.
by machinelearning 5y ago
Having had quite a bit of experience on both sides of the equation, you’re absolutely right. The very act of extrapolating how a person would perform based on a brief snapshot in time seems flawed. It is unreasonable to task the interviewer

More ›