@MengmiZhang Profile picture

Mengmi Zhang

@MengmiZhang

Nanyang assistant professor in AI and computational neuroscience @ NTU Singapore, principal scientist and PI of Deep NeuroCognitionLab @ A*star

Joined February 2012
Similar User
Victor.Kai Wang photo

@VictorKaiWang1

Mike Shou photo

@MikeShou1

Gabriel Kreiman photo

@gkreiman

Fangzhou Hong photo

@hongfz16

Xingyi Yang photo

@yxy2168

Xinchao Wang photo

@XinchaoWang3

Ravi Srinivasan photo

@ravifrancesco

Jing Liu photo

@JingLiu19624459

Zilong Zheng photo

@ZilongZheng

Yulei Niu photo

@Yulei_Niu

YUCHAO GU photo

@YuchaoGu

Caitlin Mallory photo

@sc8lin

Jay Z. Wu photo

@jayzhangjiewu

Vivek Prajapati photo

@vivekpr24

David J. Zhang photo

@JunhaoZHANG19

Mengmi Zhang Reposted

The era of post training has begun.

Yall heard it from the man himself

_jasonwei's tweet image. Yall heard it from the man himself


Exciting news! 🚨 Our lab will have two poster presentations at #NeurIPS this year! 🎉 I’ll be at the conference and would love to chat—whether it’s about science, our work, or opportunities in the lab (we’re hiring PhD positions!). Stop by and say hi! 👋


Mengmi Zhang Reposted

I don't wanna say "I told you so", but I told you so. Quote: "Ilya Sutskever, co-founder of AI labs Safe Superintelligence (SSI) and OpenAI, told Reuters recently that results from scaling up pre-training - the phase of training an AI model that uses a vast amount of unlabeled…


The video presentation of our work on Adaptive visual scene understanding in continual learning is LIVE! 🌍 youtu.be/BRRdyI9cWiY See you all in Vancouver next month!✨ #NeurIPS #AI #ContinualLearning


The video presentation of our work on generalization to biological motion perception is LIVE! 🌍 youtu.be/bZz1UZa5QLc?si… See you all in Vancouver next month!✨ #NeurIPS #AI #MachineLearning #HumanAI


🌟Imagine an AI that can continually learn to recognize new objects and build relationships in real time! Our latest #NeurIPS work on adaptive scene understanding tackles real-world complexities in dynamic environments: arxiv.org/pdf/2310.01636 #AI #ContinualLearning #NeurIPS2024

MengmiZhang's tweet image. 🌟Imagine an AI that can continually learn to recognize new objects and build relationships in real time! Our latest #NeurIPS work on adaptive scene understanding tackles real-world complexities in dynamic environments: arxiv.org/pdf/2310.01636 #AI #ContinualLearning #NeurIPS2024

📢 Can you recognize what these dots are doing? 🕺 You can—but AI can’t! Our latest work at #NeurIPS explores human and AI perception of biological motion patterns. Discover how we're bridging this gap here: arxiv.org/abs/2405.16493! #AI #Neuroscience #NeurIPS2024


Ciao! 🇮🇹 I'll be at ECCV this afternoon (Sep 29), presenting a talk on neuroscience-inspired AI at the Human-inspired Computer Vision workshop. Looking forward to seeing you there! I'm also available to chat on any other day during the conference. sites.google.com/view/hcvworksh…


Current language models are useful tools but NOT intelligent. They lack continual learning abilities, and zero-shot approaches are not the answer for catastrophic forgetting. Frustrated with #NeurIPS reviews. We need a shift in how we evaluate progress in AI. #ContinualLearning


👀 Calling all vision science researchers! Excited to announce that abstract submissions and conference registrations are now open for 16th Asia Pacific Conference on Vision in Singapore. 🌟Submit your abstracts at apcv2024.com #APCV2024 #VisionScience #Singapore 📝🔬


Mengmi Zhang Reposted

Spend 3 weeks in Woods Hole with a great NeuroAI crowd! (note that the application is open to internationals)

Applications are invited to participate in the Brains, Minds and Machines summer school [08/04/2024 -- 08/25/2024] in Woods Hole, MA. Deadline = 03/20/2024. More information here: klab.tch.harvard.edu/academia/class….

gkreiman's tweet image. Applications are invited to participate in the Brains, Minds and Machines summer school [08/04/2024 -- 08/25/2024] in Woods Hole, MA. Deadline = 03/20/2024. More information here: klab.tch.harvard.edu/academia/class….


Mengmi Zhang Reposted

My paper showing Invariant Neural Correlates for Parts of Speech in the Human Brain using Intracranial Depth Electrodes biorxiv.org/cgi/content/sh…


How can compositionality and external memory play a part in replay for continual learning? This work is in collaboration with amazing colleagues from KreimanLab @gkreiman

New ideas on using an external memory and efficient feature replays for continual learning by Talbot et al in IEEE Transactions in Neural Networks and Learning Systems, 2023. See paper here: klab.tch.harvard.edu/publications/p…

gkreiman's tweet image. New ideas on using an external memory and efficient feature replays for continual learning by Talbot et al in IEEE Transactions in Neural Networks and Learning Systems, 2023. See paper here: klab.tch.harvard.edu/publications/p…


🎓 My lab is hiring for multiple roles: full-time and part-time Ph.D. positions for August intake,2024 + Postdoc positions. If you're passionate about interdisciplinary research in AI and computational neuroscience, send your CV to me #NTUsg #ResearchJobs #PhD #Postdoc #HiringNow

MengmiZhang's tweet image. 🎓 My lab is hiring for multiple roles: full-time and part-time Ph.D. positions for August intake,2024 + Postdoc positions. If you're passionate about interdisciplinary research in AI and computational neuroscience, send your CV to me #NTUsg #ResearchJobs #PhD #Postdoc #HiringNow

Mengmi Zhang Reposted

Only academics care about generalization, industry cares about profit. When google still publishes such stuff, OpenAI directly trains on testing set without asking :) So for every industry guy, start to collect all testing case as training data now!

Ummm ... why is this a surprise? Transformers are not elixirs. Machine learning 101: gotta cover the test distribution in training! LLMs work so well because they are trained on (almost) all text distribution of tasks that we care about. That's why data quality is number 1…



Exciting news!🚀Our NeurIPS paper on "Object-centric Learning with Cyclic Walks between Parts and Whole" is now live! Learn compositionality via walks between entities in external memory and features wo decoders. 🧠🤖Paper is here: [arxiv.org/pdf/2302.08023…] #NeurIPS #AIResearch

MengmiZhang's tweet image. Exciting news!🚀Our NeurIPS paper on "Object-centric Learning with Cyclic Walks between Parts and Whole" is now live! Learn compositionality via walks between entities in external memory and features wo decoders. 🧠🤖Paper is here: [arxiv.org/pdf/2302.08023…] #NeurIPS #AIResearch

🎉 Proud to share our latest NeurIPS publication, "Decoding the Enigma" @sikarwar_ank! We've benchmarked humans and AIs on the intricate aspects of Working Memory. Check out the research here: arxiv.org/pdf/2307.10768… 🧠💡 #NeurIPS #WorkingMemory #AIResearch

MengmiZhang's tweet image. 🎉 Proud to share our latest NeurIPS publication, "Decoding the Enigma" @sikarwar_ank! We've benchmarked humans and AIs on the intricate aspects of Working Memory. Check out the research here: arxiv.org/pdf/2307.10768… 🧠💡 #NeurIPS #WorkingMemory #AIResearch

Mengmi Zhang Reposted

Introducing SignAvatars [signavatars.github.io], to the best of my knowledge, the first large-scale, multi-prompt 3D #SignLanguage motion #dataset designed to bridge the communication gap for hearing-impaired people. It is also my first take on this important problem. #CVPR👇

tolga_birdal's tweet image. Introducing SignAvatars [signavatars.github.io], to the best of my knowledge, the first large-scale, multi-prompt 3D #SignLanguage motion #dataset designed to bridge the communication gap for hearing-impaired people. It is also my first take on this important problem. #CVPR👇

Mengmi Zhang Reposted

Behold: the #ICCV2023 proceedings has been released! openaccess.thecvf.com/ICCV2023 See you in Paris!

wjscheirer's tweet image. Behold: the #ICCV2023 proceedings has been released!

openaccess.thecvf.com/ICCV2023

See you in Paris!

Excited to share that two NeurIPS papers have been accepted in our lab! Congratulations to two amazing lab members @sikarwar_ank and Ziyu Wang! Camera-ready versions with code and supp materials will be made public soon! Stay tuned! #NeurIPS2023


Loading...

Something went wrong.


Something went wrong.