doi.bio/tom_sercu

Tom Sercu

Tom Sercu is a Research Engineer and Tech Lead Manager for the protein team at Facebook AI Research in New York City. His research focuses on using artificial intelligence to impact structural biology and has covered various areas of machine learning and deep learning.

Education

Sercu holds a Bachelor of Science/Master of Science in Engineering Physics from Ghent University and a Master of Science in Data Science from New York University's Courant Institute of Mathematical Sciences.

Career

Before joining Facebook, Sercu was at IBM Research AI in the T.J. Watson Research Center. He has also worked as a Teaching Assistant for two courses at NYU: Yann LeCun's Deep Learning Course and the Programming for Data Science lab.

Research

Sercu's research interests include generative adversarial networks (GANs), semi-supervised learning, multimodal learning, and acoustic modeling in speech recognition. He has published several papers in these areas, including:

"Semi-Supervised Learning with IPM-based GANS: an Empirical Study" (NIPS Workshop, 2017)
"Dense Prediction on Sequences with Time-Dilated Convolutions for Speech Recognition" (NIPS End-to-end Learning for Speech and Audio Processing Workshop, 2016)
"Very deep multilingual convolutional neural networks for LVCSR" (Proc ICASSP, 2015)
"Multi-Frame Cross-Entropy Training for Convolutional Neural Networks in Speech Recognition" (co-authored with Neil Mallinar)
"Wasserstein Barycenter Model Ensembling" (co-authored with Pierre Dognin, Igor Melnyk, Youssef Mroueh, Jerret Ross, and Cicero Dos Santos)
"PepCVAE: Semi-Supervised Targeted Design of Antimicrobial Peptide Sequences" (co-authored with Payel Das, Kahini Wadhawan, Oscar Chang, Cicero Dos Santos, Matthew Riemer, Inkit Padhi, Vijil Chenthamarakshan, and Aleksandra Mojsilovic)
"Big-Little Net: An Efficient Multi-Scale Feature Representation for Visual and Speech Recognition" (co-authored with Chun-Fu Chen, Quanfu Fan, Neil Mallinar, and Rogerio Feris)
"Improved Image Captioning with Adversarial Semantic Alignment" (co-authored with Pierre L. Dognin, Igor Melnyk, Youssef Mroueh, and Jarret Ross)
"Regularized Kernel and Neural Sobolev Descent: Dynamic MMD Transport" (co-authored with Youssef Mroueh and Anant Raj)

Sercu has also given guest lectures on deep learning at CCNY and has taught an intro to Deep Learning course for undergraduate students in Computer Science.

Publications

Sercu's other publications include:

"Interactive Visual Exploration of Latent Space (IVELS) for peptide auto-encoder model selection" (ICLR workshop: Deep Generative Models for Highly Structured Data, 2019)
"Sobolev Descent" (AISTATS, 2019, co-authored with Youssef Mroueh and Anant Raj)
"Accelerating Antimicrobial Discovery with Controllable Deep Generative Models and Molecular Dynamics" (co-authored with Payel Das, Kahini Wadhawan, Inkit Padhi, Sebastian Gehrmann, Flaviu Cipcigan, Vijil Chenthamarakshan, Hendrik Strobelt, Cicero dos Santos, and Pin-Yu Chen)
"Sobolev Independence Criterion" (co-authored with Youssef Mroueh, Mattia Rigotti, Inkit Padhi, and Cicero Dos Santos)

https://twitter.com/TomSercu

+begin_ai

To view keyboard shortcuts, press question mark View keyboard shortcuts Tom Sercu 334 posts

See new posts Tom Sercu @TomSercu Building
@evoscaleai

Frontier AI for biology. Ex-Meta FAIR, Ex-IBM Research. Alum @NYU , @ugent . New Yorktom.sercu.meJoined February 2012 710 Following 2,074 Followers Followed by He He, Machine learning for protein engineering seminar, and 15 others you follow Posts Replies Media Tom Sercu’s posts Tom Sercu reposted halilakin @halilakin · Jun 25 The hottest new programming language is now biology 😉 #esm3 Quote Andrej Karpathy @karpathy · Jan 24, 2023 The hottest new programming language is English Tom Sercu reposted Roshan Rao @proteinrosh · Jun 25 Working on ESM3 has been the most challenging and the most rewarding part of my career. I am incredibly proud of the team we have built - y’all make it so fun to come in to work each day. Quote Thomas Hayes @THayes427 · Jun 25 Replying to @THayes427 I’m incredibly grateful to work with this amazing team. This is the most dedicated and creative team I’ve ever worked with, and I’m so excited to continue building. Please don’t hesitate to reach out if you’re interested! Tom Sercu reposted Nicholas Sofroniew @sofroniewn · Jun 25 I'm so excited to share that I've been at EvolutionaryScale, it feels like a dream come true!

As one of our first projects, we took on designing a new green fluorescent protein with ESM3, our frontier generative language model for programming biology Quote Alex Rives @alexrives · Jun 25 We have trained ESM3 and we're excited to introduce EvolutionaryScale.

ESM3 is a generative language model for programming biology. In experiments, we found ESM3 can simulate 500M years of evolution to generate new fluorescent proteins.

Read more: https://evolutionaryscale.ai/blog/esm3-release Show more

please turn into json

[AI]:

{
  "posts": [
    {
      "author": "Tom Sercu",
      "handle": "@TomSercu",
      "bio": "Building @evoscaleai - Frontier AI for biology. Ex-Meta FAIR, Ex-IBM Research. Alum @NYU, @ugent.",
      "location": "New York",
      "website": "tom.sercu.me",
      "joinedDate": "February 2012",
      "following": 710,
      "followers": 2074,
      "context": "Followed by He He, Machine learning for protein engineering seminar, and 15 others you follow",
      "reposts": [
        {
          "originalAuthor": "halilakin",
          "originalAuthorHandle": "@halilakin",
          "originalPostDate": "Jun 25",
          "content": "The hottest new programming language is now biology 😉 #esm3",
          "quote": {
            "quotedAuthor": "Andrej Karpathy",
            "quotedAuthorHandle": "@karpathy",
            "quotedContent": "The hottest new programming language is English",
            "quoteDate": "Jan 24, 2023"
          }
        },
        {
          "originalAuthor": "Roshan Rao",
          "originalAuthorHandle": "@proteinrosh",
          "originalPostDate": "Jun 25",
          "content": "Working on ESM3 has been the most challenging and the most rewarding part of my career. I am incredibly proud of the team we have built - y’all make it so fun to come in to work each day.",
          "quote": {
            "quotedAuthor": "Thomas Hayes",
            "quotedAuthorHandle": "@THayes427",
            "quotedContent": "I’m incredibly grateful to work with this amazing team. This is the most dedicated and creative team I’ve ever worked with, and I’m so excited to continue building. Please don’t hesitate to reach out if you’re interested!",
            "quoteDate": "Jun 25"
          }
        },
        {
          "originalAuthor": "Nicholas Sofroniew",
          "originalAuthorHandle": "@sofroniewn",
          "originalPostDate": "Jun 25",
          "content": "I'm so excited to share that I've been at EvolutionaryScale, it feels like a dream come true!\n\nAs one of our first projects, we took on designing a new green fluorescent protein with ESM3, our frontier generative language model for programming biology",
          "quote": {
            "quotedAuthor": "Alex Rives",
            "quotedAuthorHandle": "@alexrives",
            "quotedContent": "We have trained ESM3 and we're excited to introduce EvolutionaryScale.\n\nESM3 is a generative language model for programming biology. In experiments, we found ESM3 can simulate 500M years of evolution to generate new fluorescent proteins.\n\nRead more: https://evolutionaryscale.ai/blog/esm3-release",
            "quoteDate": "Jun 25"
          }
        }
      ]
    }
  ]
}

Tom Sercu

Tom Sercu is a Research Engineer and Tech Lead Manager for the protein team at Facebook AI Research in New York City. He works with the protein team to apply Artificial Intelligence methods to structural biology.

Education

Sercu graduated with an MS in Data Science from New York University's Courant Institute of Mathematical Sciences and holds a B.Sc./M.Sc. in Engineering Physics from Ghent University.

Career

Before joining Facebook, Sercu was at IBM Research AI in the T.J. Watson Research Center. His research has covered various areas of machine learning and deep learning, including learning representations and generative models for proteins and peptides, and unsupervised and semi-supervised learning. He has also worked on Generative Adversarial Networks (GANs), multimodal learning, and acoustic modeling in speech recognition.

Publications

Sercu has numerous publications in the fields of machine learning and AI. Some of his notable works include:

"Learning inverse folding from millions of predicted structures" (ICML 2022) with Chloe Hsu, Robert Verkuil, Jason Liu, and others.
"Biological structure and function emerge from scaling unsupervised learning to 250 million protein sequences" (Proc. Natl. Acad. Sci. USA 2021) with Alexander Rives, Joshua Meier, and others.
"Improved Mutual Information Estimation" (AAAI 2021) with Youssef Mroueh, Igor Melnyk, Pierre L. Dognin, and Jarret Ross.
"Transformer protein language models are unsupervised structure learners" (ICLR 2021) with Roshan Rao, Joshua Meier, and others.
"Sobolev Descent" (AISTATS 2019) with Youssef Mroueh and Anant Raj.
"Semi-Supervised Learning with IPM-based GANs: an Empirical Study" (NIPS Workshop 2017) with Youssef Mroueh.
"Dense Prediction on Sequences with Time-Dilated Convolutions for Speech Recognition" (NIPS End-to-end Learning for Speech and Audio Processing Workshop 2016) with Vaibhava Goel.
"Very deep multilingual convolutional neural networks for LVCSR" (Proc ICASSP 2015) with Christian Puhrsch, Brian Kingsbury, and Yann LeCun.

Tom Sercu

Tom Sercu is a Research Engineer and Tech Lead Manager for the protein team at Facebook AI Research in New York City. His work focuses on using Artificial Intelligence to advance structural biology.

Education

Sercu has an MS in Data Science from New York University's Courant Institute of Mathematical Sciences and a B.Sc./M.Sc. in Engineering Physics from Ghent University.

Career

Before joining Facebook, Sercu was at IBM Research AI in the T.J. Watson Research Center.

Research Focus

Sercu's research has covered several areas of machine learning and deep learning, including:

Learning representations and generative models for proteins and peptides
Unsupervised and semi-supervised learning with little to no labelled data
Generative Adversarial Networks (GANs) and finding a better distance metric for fast and stable training
Multimodal learning across different data modalities like images, text, and speech
Deep learning approaches to acoustic modelling in speech recognition

Publications

Sercu has numerous publications, including:

"Semi-Supervised Learning with IPM-based GANs: an Empirical Study" (NIPS Workshop: Deep Learning: Bridging Theory and Practice, 2017)
"Dense Prediction on Sequences with Time-Dilated Convolutions for Speech Recognition" (NIPS End-to-end Learning for Speech and Audio Processing Workshop, 2016)
"Very deep multilingual convolutional neural networks for LVCSR" (Proc ICASSP, 2015)
"Sobolev Descent. AISTATS" (2019)

- "Language models of protein sequences at the scale of evolution enable accurate structure prediction" (Proceedings of the National Academy of Sciences)

Tom Sercu

Education

Career

Research

"Semi-Supervised Learning with IPM-based GANS: an Empirical Study" (NIPS Workshop, 2017)
"Dense Prediction on Sequences with Time-Dilated Convolutions for Speech Recognition" (NIPS End-to-end Learning for Speech and Audio Processing Workshop, 2016)
"Very deep multilingual convolutional neural networks for LVCSR" (Proc ICASSP, 2015)
"Multi-Frame Cross-Entropy Training for Convolutional Neural Networks in Speech Recognition" (co-authored with Neil Mallinar)
"Wasserstein Barycenter Model Ensembling" (co-authored with Pierre Dognin, Igor Melnyk, Youssef Mroueh, Jerret Ross, and Cicero Dos Santos)
"PepCVAE: Semi-Supervised Targeted Design of Antimicrobial Peptide Sequences" (co-authored with Payel Das, Kahini Wadhawan, Oscar Chang, Cicero Dos Santos, Matthew Riemer, Inkit Padhi, Vijil Chenthamarakshan, and Aleksandra Mojsilovic)
"Big-Little Net: An Efficient Multi-Scale Feature Representation for Visual and Speech Recognition" (co-authored with Chun-Fu Chen, Quanfu Fan, Neil Mallinar, and Rogerio Feris)
"Improved Image Captioning with Adversarial Semantic Alignment" (co-authored with Pierre L. Dognin, Igor Melnyk, Youssef Mroueh, and Jarret Ross)
"Regularized Kernel and Neural Sobolev Descent: Dynamic MMD Transport" (co-authored with Youssef Mroueh and Anant Raj)

Sercu has also given guest lectures on deep learning at CCNY and has taught an intro to Deep Learning course for undergraduate students in Computer Science.

Publications

Sercu's other publications include:

"Interactive Visual Exploration of Latent Space (IVELS) for peptide auto-encoder model selection" (ICLR workshop: Deep Generative Models for Highly Structured Data, 2019)
"Sobolev Descent" (AISTATS, 2019, co-authored with Youssef Mroueh and Anant Raj)
"Accelerating Antimicrobial Discovery with Controllable Deep Generative Models and Molecular Dynamics" (co-authored with Payel Das, Kahini Wadhawan, Inkit Padhi, Sebastian Gehrmann, Flaviu Cipcigan, Vijil Chenthamarakshan, Hendrik Strobelt, Cicero dos Santos, and Pin-Yu Chen)
"Sobolev Independence Criterion" (co-authored with Youssef Mroueh, Mattia Rigotti, Inkit Padhi, and Cicero Dos Santos)

Youtube Videos

Youtube Title: June 4, 2023

Youtube Link: link

Youtube Channel Name: Tom Sercu

Youtube Channel Link: https://www.youtube.com/@tomsercu6963

June 4, 2023

Youtube Title: May 30, 2023

Youtube Link: link

Youtube Channel Name: Tom Sercu

Youtube Channel Link: https://www.youtube.com/@tomsercu6963

May 30, 2023

Youtube Title: NYU Data Science in New York City

Youtube Link: link

Youtube Channel Name: NYU Data Science

Youtube Channel Link: https://www.youtube.com/@nyudatascience7498

NYU Data Science in New York City

Youtube Title: C3.ai CEO Tom Siebel on the future of AI

Youtube Link: link

Youtube Channel Name: CNBC Television

Youtube Channel Link: https://www.youtube.com/@CNBCtelevision

C3.ai CEO Tom Siebel on the future of AI

Youtube Title: Kašubovci: Chaka Chaka (Cover)

Youtube Link: link

Youtube Channel Name: tv kasino

Youtube Channel Link: https://www.youtube.com/@tvkasino

Kašubovci: Chaka Chaka (Cover)

Youtube Title: C3 AI CEO Tom Siebel on Fox Business

Youtube Link: link

Youtube Channel Name: C3 AI

Youtube Channel Link: https://www.youtube.com/@C3AI

C3 AI CEO Tom Siebel on Fox Business

Youtube Title: Sobolev Independence Criterion: Non-Linear Feature Selection with False Discovery Control.

Youtube Link: link

Youtube Channel Name: MITCBMM

Youtube Channel Link: https://www.youtube.com/@MITCBMM

Sobolev Independence Criterion: Non-Linear Feature Selection with False Discovery Control.

Youtube Title: Generative AI Changes Everything | Tom Siebel Keynote

Youtube Link: link

Youtube Channel Name: C3 AI

Youtube Channel Link: https://www.youtube.com/@C3AI

Generative AI Changes Everything | Tom Siebel Keynote

Youtube Title: The Digital Journey, Tom Siebel | C3 Transform 2022

Youtube Link: link

Youtube Channel Name: C3 AI

Youtube Channel Link: https://www.youtube.com/@C3AI

The Digital Journey, Tom Siebel | C3 Transform 2022

Youtube Title: Rihanna - Lift Me Up (Wakanda Forever Lyric Video)

Youtube Link: link

Youtube Channel Name: Rihanna

Youtube Channel Link: https://www.youtube.com/channel/UCcgqSM4YEo5vVQpqwN-MaNw

Rihanna - Lift Me Up (Wakanda Forever Lyric Video)

Youtube Title: mvadrill

Youtube Link: link

Youtube Channel Name: jdeserto

Youtube Channel Link: https://www.youtube.com/@jdeserto

mvadrill

Youtube Title: YouTube Automation with AI - FULL COURSE (10+ Hours)

Youtube Link: link

Youtube Channel Name: AI Guy

Youtube Channel Link: https://www.youtube.com/@ai_guy

YouTube Automation with AI - FULL COURSE (10+ Hours)

Youtube Title: 528 Hz Miracle - Tom Soltron - Solfeggio Dynamic Meditation

Youtube Link: link

Youtube Channel Name: TheToneOfLife

Youtube Channel Link: https://www.youtube.com/@TheToneoflife

528 Hz Miracle - Tom Soltron - Solfeggio Dynamic Meditation

Youtube Title: C3 AI CEO Tom Siebel on Unprecedented AI Demand | Bloomberg TV

Youtube Link: link

Youtube Channel Name: C3 AI

Youtube Channel Link: https://www.youtube.com/@C3AI

C3 AI CEO Tom Siebel on Unprecedented AI Demand | Bloomberg TV

Youtube Title: Pink Floyd - " MOTHER " The Wall 1980

Youtube Link: link

Youtube Channel Name: HDPinkFloyd

Youtube Channel Link: https://www.youtube.com/@HDPinkFloyd

Pink Floyd - " MOTHER " The Wall 1980

Youtube Title: Grace To Grace - Hillsong Worship

Youtube Link: link

Youtube Channel Name: Hillsong Worship

Youtube Channel Link: https://www.youtube.com/channel/UC4q12NoPNySbVqwpw4iO5Vg

Grace To Grace - Hillsong Worship

Youtube Title: Thomas Dolby: "Love Is a Loaded Pistol"

Youtube Link: link

Youtube Channel Name: TED

Youtube Channel Link: https://www.youtube.com/@TED

Thomas Dolby: "Love Is a Loaded Pistol"

Youtube Title: Machine Learning for Computational Biology (MLCB) Virtual Meeting 2020

Youtube Link: link

Youtube Channel Name: abk2001

Youtube Channel Link: https://www.youtube.com/@abk2001

Machine Learning for Computational Biology (MLCB) Virtual Meeting 2020

Youtube Title: We will see massive innovation in artificial intelligence | C3 AI CEO Tom Siebel on Fox Business

Youtube Link: link

Youtube Channel Name: C3 AI

Youtube Channel Link: https://www.youtube.com/@C3AI

We will see massive innovation in artificial intelligence | C3 AI CEO Tom Siebel on Fox Business

Youtube Videos

Youtube Title: June 4, 2023

Youtube Link: link

Youtube Channel Name: Tom Sercu

Youtube Channel Link: https://www.youtube.com/@tomsercu6963

June 4, 2023

Youtube Title: May 30, 2023

Youtube Link: link

Youtube Channel Name: Tom Sercu

Youtube Channel Link: https://www.youtube.com/@tomsercu6963

May 30, 2023

Youtube Title: NYU Data Science in New York City

Youtube Link: link

Youtube Channel Name: NYU Data Science

Youtube Channel Link: https://www.youtube.com/@nyudatascience7498

NYU Data Science in New York City

Youtube Title: C3.ai CEO Tom Siebel on the future of AI

Youtube Link: link

Youtube Channel Name: CNBC Television

Youtube Channel Link: https://www.youtube.com/@CNBCtelevision

C3.ai CEO Tom Siebel on the future of AI

Youtube Title: Kašubovci: Chaka Chaka (Cover)

Youtube Link: link

Youtube Channel Name: tv kasino

Youtube Channel Link: https://www.youtube.com/@tvkasino

Kašubovci: Chaka Chaka (Cover)

Youtube Title: C3 AI CEO Tom Siebel on Fox Business

Youtube Link: link

Youtube Channel Name: C3 AI

Youtube Channel Link: https://www.youtube.com/@C3AI

C3 AI CEO Tom Siebel on Fox Business

Youtube Title: Sobolev Independence Criterion: Non-Linear Feature Selection with False Discovery Control.

Youtube Link: link

Youtube Channel Name: MITCBMM

Youtube Channel Link: https://www.youtube.com/@MITCBMM

Sobolev Independence Criterion: Non-Linear Feature Selection with False Discovery Control.

Youtube Title: Generative AI Changes Everything | Tom Siebel Keynote

Youtube Link: link

Youtube Channel Name: C3 AI

Youtube Channel Link: https://www.youtube.com/@C3AI

Generative AI Changes Everything | Tom Siebel Keynote

Youtube Title: The Digital Journey, Tom Siebel | C3 Transform 2022

Youtube Link: link

Youtube Channel Name: C3 AI

Youtube Channel Link: https://www.youtube.com/@C3AI

The Digital Journey, Tom Siebel | C3 Transform 2022

Youtube Title: Rihanna - Lift Me Up (Wakanda Forever Lyric Video)

Youtube Link: link

Youtube Channel Name: Rihanna

Youtube Channel Link: https://www.youtube.com/channel/UCcgqSM4YEo5vVQpqwN-MaNw

Rihanna - Lift Me Up (Wakanda Forever Lyric Video)

Youtube Title: mvadrill

Youtube Link: link

Youtube Channel Name: jdeserto

Youtube Channel Link: https://www.youtube.com/@jdeserto

mvadrill

Youtube Title: YouTube Automation with AI - FULL COURSE (10+ Hours)

Youtube Link: link

Youtube Channel Name: AI Guy

Youtube Channel Link: https://www.youtube.com/@ai_guy

YouTube Automation with AI - FULL COURSE (10+ Hours)

Youtube Title: 528 Hz Miracle - Tom Soltron - Solfeggio Dynamic Meditation

Youtube Link: link

Youtube Channel Name: TheToneOfLife

Youtube Channel Link: https://www.youtube.com/@TheToneoflife

528 Hz Miracle - Tom Soltron - Solfeggio Dynamic Meditation

Youtube Title: C3 AI CEO Tom Siebel on Unprecedented AI Demand | Bloomberg TV

Youtube Link: link

Youtube Channel Name: C3 AI

Youtube Channel Link: https://www.youtube.com/@C3AI

C3 AI CEO Tom Siebel on Unprecedented AI Demand | Bloomberg TV

Youtube Title: Pink Floyd - " MOTHER " The Wall 1980

Youtube Link: link

Youtube Channel Name: HDPinkFloyd

Youtube Channel Link: https://www.youtube.com/@HDPinkFloyd

Pink Floyd - " MOTHER " The Wall 1980

Youtube Title: Grace To Grace - Hillsong Worship

Youtube Link: link

Youtube Channel Name: Hillsong Worship

Youtube Channel Link: https://www.youtube.com/channel/UC4q12NoPNySbVqwpw4iO5Vg

Grace To Grace - Hillsong Worship

Youtube Title: Thomas Dolby: "Love Is a Loaded Pistol"

Youtube Link: link

Youtube Channel Name: TED

Youtube Channel Link: https://www.youtube.com/@TED

Thomas Dolby: "Love Is a Loaded Pistol"

Youtube Title: Machine Learning for Computational Biology (MLCB) Virtual Meeting 2020

Youtube Link: link

Youtube Channel Name: abk2001

Youtube Channel Link: https://www.youtube.com/@abk2001

Machine Learning for Computational Biology (MLCB) Virtual Meeting 2020

Youtube Title: We will see massive innovation in artificial intelligence | C3 AI CEO Tom Siebel on Fox Business

Youtube Link: link

Youtube Channel Name: C3 AI

Youtube Channel Link: https://www.youtube.com/@C3AI

We will see massive innovation in artificial intelligence | C3 AI CEO Tom Siebel on Fox Business