xlvector

researcher on Recommender System.

[Jason Wei @_jasonwei]

Very excited to finally share OpenAI's "deep research" model, which achieves twice the score of o3-mini on Humanity's Last Exam, and can even perform some tasks that would take PhD experts 10+ hours to do!

A few thoughts on the implications: Deep research can be seen as a new interface for the internet, in addition to being an incredible agent. The current way that we humans use the internet is quite clunky--it takes time to search and click around, we're limited by time and attention, and we don't know the full range of websites out there.

AI, on the other hand, never gets tired, can browse many websites at a time and has nearly unlimited world knowledge. With a deep research model that browses the web so well, we are entering a world where any information that would take a human hours to compile, can be synthesized by AI for you in a few minutes.

We now have the luxury of:
- Instead of combing through hundreds of reviews to decide on a hotel, restaurant, or product, AI will synthesize all the relevant info for us
- When we travel, we'll get the best tips from all corners of the web, even ones from niche forums
- For any new scientific idea, we can quickly get a list of all the relevant literature, related ideas, and people to connect with

It's basically like a version of the internet personalized for what you want to know. This paradigm will be so powerful that in the future, navigating the internet manually via a browser will be "old-school", like performing arithmetic calculations by hand instead of using a calculator.

[xlvector @xlvector]

@topocheng 回来了？

[xlvector @xlvector]

If your product s#@ks, marketing won't do s#!t - Steve Jobs https://t.co/8b2Wq8YrSW 来自 @YouTube

[xlvector @xlvector]

I think we found the newest Mortal Kombat Level. https://t.co/1ijOxL0iWs 来自 @YouTube

[xlvector @xlvector]

The Big Misconception About Electricity https://t.co/9izTkcImHe 来自 @YouTube

[Yann LeCun @ylecun]

Self-Supervised Learning making strides in speech recognition.

Wav2Vec 2.0 from FAIR uses a kind of contrastive SSL for pre-training.

This is the first time an SSL system reaches the very best results on a number of different ASR tasks.

https://t.co/pqu1E320f4
1/N

[xlvector @xlvector]

有对头条西瓜算法有兴趣的同学可以联系我

[xlvector @xlvector]

我写了新文章《word2vec/lstm on mxnet with NCE loss》（ 分享自 @jianshucom ） https://t.co/tEKo6kGWaV

[xlvector @xlvector]

An overview of gradient descent optimization algorithms » https://t.co/TZxyMT4TxY #optimization #deep learning #sgd

[xlvector @xlvector]

Exploitation and Exploration http://t.co/wbSCrfy0

[xlvector @xlvector]

推荐系统实践勘误 http://t.co/nmUJeP50

[xlvector @xlvector]

Netflix推荐系统（第二部分）—— 排序 http://t.co/NpO3zUwu

[xlvector @xlvector]

Min-Hash和推荐系统 http://t.co/0Qaxq91M

[xlvector @xlvector]

郑欣然：一名警察法 轮 功修炼者的故事 http://t.co/8aSd8yzV

[xlvector @xlvector]

@DrButner You can find it on the bottom of http://t.co/bviyrUyf

[xlvector @xlvector]

research papers about twitter in reculike http://t.co/KAdnYf0t

[xlvector @xlvector]

Recommend paper "Adaptive Bootstrapping of Recommender Systems Using Decision Trees" http://t.co/xu4Vy0mi

[xlvector @xlvector]

Power to the people: exploring neighbourhood formations in social recommender system - RecULike http://t.co/5KMqzqSc

[xlvector @xlvector]

How Hulu’s ‘Steamboat’ ads recommend new shows to users http://t.co/NcRIUtGe

[xlvector @xlvector]

《推荐系统实践》样章：如何利用用户标签数据（二） http://t.co/3jKIbxMT