Explore trending content on Musk Viewer
#WWERaw
• 129274 Tweets
ブロック
• 127028 Tweets
Bengals
• 105672 Tweets
Jayden Daniels
• 79166 Tweets
うまい棒
• 75435 Tweets
連休明け
• 65674 Tweets
Terry
• 63474 Tweets
GFRIEND
• 54593 Tweets
呂布カルマ
• 50388 Tweets
Jey Uso
• 46462 Tweets
VOTA X KARIME
• 45424 Tweets
#沼におぼれる
• 20529 Tweets
ラッパー
• 19828 Tweets
lorde
• 19794 Tweets
イーロン
• 17466 Tweets
tini
• 16843 Tweets
インプレゾンビ
• 16638 Tweets
SATUIN SEMANGAT Pasti RIDO
• 16572 Tweets
SATUKAN Energi POSITIF Kita
• 15851 Tweets
クロススカウト
• 13758 Tweets
コレクション缶バッジ
• 11564 Tweets
Pinned Tweet
Thrilled to release o1-mini, a model near and dear to my heart 💙. o1-mini is an efficient model in the o1 series that’s super performant in STEM reasoning, especially math and coding. I can’t wait to see what you all build with o1-mini!!
17
26
303
Excited to be speaking at the Stanford Graph Learning Workshop!
Free registration:
More info:
Excited to announce 2nd Stanford Graph Learning Workshop on Wed Sept 28th with leaders from academia and industry to showcase recent advances of Graph Representation Learning across a wide range of applications. Program & free registration:
4
159
752
0
34
294
📣It is time to rethink graph databases in the era of GNNs and neural reasoners. In the new work with
@michael_galkin
,
@michaelcochez
,
@zhu_zhaocheng
, and
@jure
we explore the concept of Neural Graph Databases (NGDBs)!
📜
🌐
🧵1/n
3
46
180
做了一点微小的贡献🀄️
Our new GPT-4 Turbo is now available to paid ChatGPT users. We’ve improved capabilities in writing, math, logical reasoning, and coding.
Source:
657
1K
7K
12
3
177
💕
3
7
156
I have been working on 4o mini 🚀, our most intelligent and efficient model. Hope it can unlock a lot more downstream applications, check it out!!
Introducing GPT-4o mini! It’s our most intelligent and affordable small model, available today in the API. GPT-4o mini is significantly smarter and cheaper than GPT-3.5 Turbo.
164
616
3K
15
5
134
Excited to share our
#ICLR2020
paper and code for Query2box, a multi-hop reasoning framework on knowledge graphs. We design box embeddings to answer complex logical queries with conjunction and disjunction. Joint work with
@weihua916
and
@jure
.
2
22
116
🚨Glad to share that Combiner is accepted to
#NeurIPS2021
as a spotlight! Combiner achieves full attention with subquadratic complexity. Joint work with amazing folks
@hanjundai
@ZihangDai
@mengjiao_yang
@jure
Dale Schuurmans
@daibond_alpha
@GoogleAI
@StanfordAILab
img from wiki
1
18
75
New
#NeurIPS2020
paper on multi-hop reasoning on knowledge graphs! joint work with
@jure
@StanfordAILab
Paper:
Website:
Code (BetaE, Query2box, GQE):
1
10
55
Come join us for the tutorial @ Learning on Graph: Complex Reasoning over Relational Database, we will present SMORE + Scallop!
@StanfordAILab
@CIS_Penn
@GoogleAI
@LogConference
Time: 9-1030am PT, Dec. 10
Webpage:
(Attendees will get a virtual cookie👇)
0
14
47
I will be flying to
#NeurIPS2023
tomorrow ✈️ and stay until Saturday. DM me if you want to chat about
#LLM
, reasoning and beyond!
1
0
40
o1-mini and preview is
#1
on math!
1
0
40
Simply would like to mark this historic moment. The past five days have been a roller coaster. Happy early thanksgiving! 🦃❤️☺️
We have reached an agreement in principle for Sam Altman to return to OpenAI as CEO with a new initial board of Bret Taylor (Chair), Larry Summers, and Adam D'Angelo.
We are collaborating to figure out the details. Thank you so much for your patience through this.
6K
13K
66K
0
0
33
Check out the clip for a preview of behind-the-scene stories of 🍓
2
0
32
Ordered some New Orleans chicken wings before heading to New Orleans
I will be flying to
#NeurIPS2023
tomorrow ✈️ and stay until Saturday. DM me if you want to chat about
#LLM
, reasoning and beyond!
1
0
40
1
0
30
Come check our
#uai
paper on task inference for meta-reinforcement learning. We propose OCEAN, an online task inference framework that models tasks with global and local context variables. Joint work with
@AnimaAnandkumar
@animesh_garg
@yukez
@jure
. 👉
1
4
26
@legit_rumors
@OpenAIDevs
- OpenAI o1-mini is optimized for STEM applications at all stages of training & data. It has limitations of world knowledge. Check our research blog post for more details.
- We are working on adding more knowledge. Stay tuned for next version of o1-mini!
0
2
26
Check out our new blog post for a summarization and outlook of the graph ML & geometric deep learning in 2022 and 2023. Happy new year! 🧨☃️🎄
👉 link:
🎄It's 2023! In a new post, we provide an overview of Graph ML and its subfields (and hypothesize for '23), eg Generative Models, Physics, PDEs, Theory, KGs, Algorithmic Reasoning, and more!
With
@ren_hongyu
@zhu_zhaocheng
@chrsmrrs
and
@jo_brandstetter
1
115
425
0
2
25
Happy to release DRAGON, a foundation model joint trained over texts and knowledge bases.
Check out GreaseLM too for the FUSION layer at the core!
Excited to share our
#NeurIPS2022
paper “DRAGON: Deep Bidirectional Language-Knowledge Graph Pretraining”, w/ the amazing
@ABosselut
@ren_hongyu
@xikun_zhang_
@chrmanning
@percyliang
@jure
@StanfordAILab
!
Paper:
Code:
🧵👇[1/6]
4
74
289
0
2
23
Check out our new scalable KG embedding framework! It supports super efficient training of single/multi-hop algorithms on extremely large graphs (~86m nodes), including GQE, Query2box, BetaE, TransE, RotatE, DistMult, ComplEx and so on!🥳
Excited to share our collaboration with
@GoogleAI
: SMORE is a scalable knowledge graph completion and multi-hop reasoning system that scales to hundreds of millions of entities and relations.
@ren_hongyu
,
@hanjundai
, et al.
2
76
403
0
5
23
👉Check out Stanford Graph Learning workshop on Sep. 16! Super excited to present recent advances on Knowledge Graphs. Free registration at: , it will be live streamed!
Stanford is proud to bring together leaders from academia&industry to showcase advances in Graph Neural Networks. Program includes applications, frameworks and industry panels on challenges of graph-based machine learning models. Register at:
11
221
897
0
3
23
New
#NeurIPS2020
paper on Graph Neural Nets
#GNN
, Representation Learning, Robustness!
"Graph Information Bottleneck" w/
@tailintalent
, Pan Li,
@jure
@StanfordAILab
@PurdueCS
Website:
Paper:
Code:
(1/n)
1
7
20
@dvyio
@OpenAIDevs
OpenAI o1-mini is optimized for STEM applications at all stages of training & data. It has limitations of world knowledge. Check our research blog post for more details.
0
1
19
👉Check out our new work QA-GNN, a new commonsense reasoning model using language models GNNs, and knowledge graphs.
#NAACL2021
Excited to share our
#NAACL2021
paper "QA-GNN: Reasoning with Language Models and Knowledge Graphs for Question Answering"! Joint work with
@ren_hongyu
@ABosselut
@percyliang
@jure
@StanfordAILab
Paper:
Github:
Thread below [1/6]
1
48
187
0
6
19
🤓The project took us months of conceptualization and refinement, and we are glad to finally release it – look forward to hearing both from Graph ML and Database folks!
Check out the blogpost from
@michael_galkin
for more details :)
10/10, n=10
1
4
17
Check out Connection Subgraph Reasoner accepted at
#NeurIPS2022
, a novel pretraining subgraph-level objective for few-shot KG link prediction.
Paper:
Code:
Excited to share our
#NeurIPS2022
paper: Few-shot Relational Reasoning via Connection
Subgraph Pretraining! . We propose to pretrain on subgraph matching for few-shot relational reasoning tasks. 🧵
Joint work with
@ren_hongyu
and
@jure
!
@StanfordAILab
4
32
218
0
7
17
@InverseAGI
@OpenAIDevs
OpenAI o1-mini is optimized for STEM applications at all stages of training & data. It has limitations of world knowledge. Check our research blog post for more details.
0
3
16
How do we answer complex queries in an inductive setting? Check out this new
#NeurIPS2022
paper with a novel method and a set of benchmarks for inductive multi-hop reasoning. This guy makes cool gifs!!
Our new
#NeurIPS2022
work on inductive query answering - new nodes at inference, complex queries over incomplete graphs, no node features, seems impossible? 🤔
With
@zhu_zhaocheng
@ren_hongyu
@tangjianpku
Paper:
Code:
🧵1/10
2
9
58
0
4
16
Come check our BetaE paper today starting 21pm PT at Town E1 - Spot B1!
@jure
New
#NeurIPS2020
paper on multi-hop reasoning on knowledge graphs! joint work with
@jure
@StanfordAILab
Paper:
Website:
Code (BetaE, Query2box, GQE):
1
10
55
0
2
15
📢Excited to give a talk about SMORE, the first scalable framework that supports link prediction and multi-hop reasoning over massive knowledge graphs!
Code:
Tuesday 10am ET Room 206
#KDD2022
0
0
13
Join the AMA session at 10am PT today! I will answer your questions.
We’re hosting an AMA for developers from 10–11 AM PT today. Reply to this thread with any questions and the OpenAI o1 team will answer as many as they can.
483
129
1K
0
1
12
☺️
o1-mini is the most surprising research result i've seen in the past year
obviously i cannot spill the secret, but a small model getting >60% on AIME math competition is so good that it's hard to believe
congrats
@ren_hongyu
@shengjia_zhao
for the great work!
35
94
2K
0
0
12
Explain superconductors! 🤓
We’re rolling out a bunch of small updates to improve the ChatGPT experience. Shipping over the next week:
1. Prompt examples: A blank page can be intimidating. At the beginning of a new chat, you’ll now see examples to help you get started.
2. Suggested replies: Go deeper with
581
1K
6K
0
0
12
Why do we need NGDBs and what do graph DBs lack? The biggest motivation is incompleteness - symbolic SPARQL/Cypher-like engines can’t cope with incomplete graphs at scale. Neural graph reasoning, however, is already mature enough to work in large and noisy incomplete graphs.
2/n
1
2
10
Broadly, NGDBs are equipped to answer both “what is there?” and “what is missing?” queries whereas standard graph DBs are limited to traversal-only scenarios assuming the graph is complete.
4/n
1
1
9
✍️Check out our github where we collect all relevant references, feel free to send PRs - we welcome all contributions :)
We’ll be updating other materials on the project website .
9/n
1
1
9
Absolutely stunning! 💫
Introducing Sora, our text-to-video model.
Sora can create videos of up to 60 seconds featuring highly detailed scenes, complex camera motion, and multiple characters with vibrant emotions.
Prompt: “Beautiful, snowy
10K
32K
138K
0
0
8
@felixchin1
@OpenAIDevs
The weekly rate limit is 50 for o1-mini. We are working to increase those rates and enable ChatGPT to automatically choose the right model for a given prompt!
1
0
8
GPT4 is smarter!!
0
0
8
What are NGDBs? While their architecture might look similar to traditional DBs with Storage and Query Engine, the essential difference is in ditching symbolic edge traversal and answering queries in the latent space (including logical operators).
3/n
1
1
8
It was an extremely wild and rewarding journey to train and babysit the model from random init to the final ckpt with the best team possible. Had sooo much fun experiencing all the TOO GOOD TO BE TRUE moments with
@shengjia_zhao
@_kevinlu
@Eric_Wallace_
@LiamFedus
@_aidan_clark_
!
0
0
7
Thanks to GNNs, NGDBs can answer either the entire query graph at once or execute them sequentially. We don’t need any explicit indexes - the latent space formed by a neural encoder *is* the single uniform index.
5/n
1
2
7
Finally, we outline challenges and open problems for NGDBs. Lots of cool stuff to work on! (especially if you are in existential crisis with GPT4, in fact how to design LLM interface for NGDB and how to let NGDBs help compress and accelerate LLMs are very promising)
#GPT4
8/n
1
1
6
What’s the difference between NGDBs and vector DBs?
Vector DBs are fast and encoder-independent consuming embeddings from all kinds of models. Though they are limited to a few distance functions and lack query answering capabilities, they fit well into the NGDB blueprint!
6/n
1
1
6
In the NGDB framework, we create a taxonomy and survey 40+ neural graph reasoning models that can potentially serve as Neural Query Engines under 3 main categories: Graphs (theory and expressiveness), Modeling (graph learning), and Queries (what can we answer).
7/n
1
1
6
@actualrealyorth
@OpenAIDevs
We are looking into this and building better infra to support our api customers and developers.
1
0
5
Check out OGB for large-scale graph ML datasets!
Super excited to share Open Graph Benchmark (OGB)! OGB provides large-scale, diverse graph datasets to catalyze graph ML research. The datasets are easily accessible via OGB Python package with unified evaluation protocols and public leaderboards.
Paper:
2
112
408
0
0
4
amazing video!
“how many r’s in strawberry?”
I had to ask this to demo our new model o1-preview 😎
LLMs process text at a subword level. A question that requires understanding the notion of both character and word confuses them.
OpenAI o1-preview "thinks harder" to avoid mistakes.
5
7
93
0
0
4
(2/x) The key idea is to factorize the conditional expectation formulation of attention in a structured way. Each token can attend to all the other tokens either by direct attention, or indirectly via a "proxy" that summarizes a local region.
Paper link:
1
2
4
@weihua916
@jure
We show that embedding queries as hyper-rectangles (box) is a natural way to handle relational projection and conjunction. We also analyze the hardness of KG embeddings handling disjunction and provide a clean solution using disjunctive normal form.
1
1
4
We embed queries as Beta distributions. For conjunctions, we take weighted product of the PDF of the Beta embeddings of input queries. For negation, we calculate the reciprocal of the parameters of the input so that the high density region will be become low and vice versa.
1
0
3
@tailin_wu
and I will present Graph Information Bottleneck from 9pm to 11pm PT today!
#NeurIPS2020
Come chat with us on graphs, representation learning, GNNs.
Website:
Gathertown:
@jure
@StanfordAILab
@PanLi90769257
@PurdueCS
New
#NeurIPS2020
paper on Graph Neural Nets
#GNN
, Representation Learning, Robustness!
"Graph Information Bottleneck" w/
@tailintalent
, Pan Li,
@jure
@StanfordAILab
@PurdueCS
Website:
Paper:
Code:
(1/n)
1
7
20
0
2
3
Check out query2box, joint work with
@weihua916
and
@jure
. Query2box performs multi-hop logical reasoning on knowledge graphs.
#ICLR2020
#iclr
Talk:
Website:
@weihua916
and I will present the work at 10 pm Tue. and 1pm Wed. PT.
0
0
3
(5/x) It's a direct drop-in replacement of the attention layers and we will release the implementation soon!
0
0
2
0
0
2
The embedding also captures the uncertainty of a query, measured by the number of answers it has. We found a natural connection between the entropy of the Beta embedding and # answers, without any training needed to enforce this correlation.
0
0
2
(3/x) We propose six instantiations of Combiner idea with different factorization patterns. And we show we can take a prior sparse Transformer X and make it Combiner-X with full attention capacity while keeping the same asymptotic complexity!
1
0
2
poster session starting soon at 7pm PT.
0
0
2
The task is answering first-order logic queries (with existential, conjunction, disjunction and negation) on incomplete KGs. Key insight is to embed the queries and entities, and reason in the embedding space. Previous methods (GQE, Q2B..) cannot handle negation/set complement.
1
0
2
(4/x) We evaluate Combiner on autoregressive / bidirectional sequence modeling across images and texts. Combiner achieves much better performance across a wide range of tasks and is super scalable.
1
0
2
The code of graph information bottleneck (GIB) can be found at , we used the fantastic Pytorch Geometric library by
@rusty1s
! Feel free to check it out!
(5/n)
1
0
1
The model achieves strong robustness against adversarial attacks on both node features and graph structure!
(4/n)
1
0
1
0
0
1
0
0
1
How do generative models generalize and memorize? Check out the blog for a systematic framework!
If all training images for a GAN/VAE/PixelCNN have 2 objects, will they only generate images with 2 objects? If trained on (🔵,💙,🔴), will they also generate ❤️? Find out in
@shengjia_zhao
's blog post on generalization and bias for generative models.
👉
1
134
515
0
0
1