Andrew Campbell

Research Scientist, Google DeepMind. Previous: @Xaira_Thera, PhD @oxcsml

[Andrew Campbell @AndrewC_ML]

@BunchNumbers @iclr_conf That sounds great, please reach out to the email on my website and we can organize :)

[Andrew Campbell @AndrewC_ML]

I will be attending @iclr_conf this week, please reach out if you'd like to chat about generative models or AI for science!

[Andrew Campbell @AndrewC_ML]

Very excited to share our preprint: Self-Speculative Masked Diffusions

We speed up sampling of masked diffusion models by ~2x by using speculative sampling and a hybrid non-causal / causal transformer

https://t.co/6e37sx8Cbu

w/ @ValentinDeBort1 @thjashin @ArnaudDoucet1

[Frank Noe @FrankNoeBerlin]

Super excited to preprint our work on developing a Biomolecular Emulator (BioEmu): Scalable emulation of protein equilibrium ensembles with generative deep learning from @MSFTResearch AI for Science.
#ML #AI #NeuralNetworks #Biology #AI4Science

https://t.co/yzOy6tAoPv

[Sulin Liu @su_lin_liu]

Discrete generative models use denoisers for generation, but they can slip up. What if generation *isn’t only* about denoising?🤔

Introducing DDPD: Discrete Diffusion with Planned Denoising🤗🧵(1/11)

w/ @junonam_ @AndrewC_ML @HannesStaerk @xuyilun2 Tommi Jaakkola @RGBLabMIT

[Andrew Campbell @AndrewC_ML]

@FrankNoeBerlin Thanks Frank! This is a choice we can make at inference time, we can sample with 'low stochasticity' where tokens commit or with 'high stochasticity' where tokens can switch back to mask (see gif). This enables the model to correct mistakes during generation.

[Andrew Campbell @AndrewC_ML]

New paper: how to do flow matching on discrete data.

Flows give a simple generative framework and better performance than discrete diffusion models.

Discrete flows are easily combined with continuous flow matching for multimodal models.

https://t.co/7eQNl1myrS

A thread 1/7

[Jason Yim @json_yim]

Combining discrete and continuous data is an important capability for generative models. To address this for protein design, we introduce Multiflow, a generative model for structure and sequence generation.
Preprint: https://t.co/wuj9l5sTLc
Code: https://t.co/IwIoC74Odm
1/8

[Andrew Campbell @AndrewC_ML]

@json_yim @valence_ai Big thanks to amazing co-lead @json_yim and our advisors @BarzilayRegina , @tom_rainforth , Tommi Jaakkola.
7/7

[Andrew Campbell @AndrewC_ML]

@json_yim We are giving a talk about the work this Tuesday 11am EST/4pm GMT @valence_ai https://t.co/d1h30l9lK6

Code for the pure discrete model: https://t.co/4zeTSX1Vv1

Code for protein co-design experiments: https://t.co/ZdOhqUOaCP
6/7

[Andrew Campbell @AndrewC_ML]

How can we apply diffusion models to data with varying dimensionality? We use jump diffusions to simultaneously generate the size and state values for varying size data e.g. molecules

https://t.co/99SvKR0NZs

w/ @willarvey @wh1lo @ValentinDeBort1 @tom_rainforth @ArnaudDoucet1

[Andrew Campbell @AndrewC_ML]

@JoeJBenton @ValentinDeBort1 @tom_rainforth @GeorgeDeligian9 @ArnaudDoucet1 Code and models available at https://t.co/uHBMFs6n9g

[Andrew Campbell @AndrewC_ML]

📰A Continuous Time Framework for Discrete Denoising Models

Operating in continuous time gives us higher performance generative samplers and error bounds on discrete spaces.

https://t.co/LwiSj0rJa7

w/ @JoeJBenton @ValentinDeBort1 @tom_rainforth @GeorgeDeligian9 @ArnaudDoucet1

[Andrew Campbell @AndrewC_ML]

I have written a blog post describing our use of Reinforcement Learning to create an online objective for sequential Variational Inference
🌐 https://t.co/es1zarYQX5
📰 https://t.co/TeJEdpYuE0

[Arnaud Doucet @ArnaudDoucet1]

Online Variational Filtering and Parameter Learning https://t.co/572qQ9ZZsv with @AndrewC_ML @YuyangShi0 & @tom_rainforth accepted as Oral #NeurIPS2021.
We exploit forward Bellman recursions for the ELBO & its gradients in state-space models and RL ideas.