Mr. Wenhao Gao (22 December 2021)

Topic: Amortized Tree Generation for Bottom-up Synthesis Planning and Synthesizable Molecular Design.

Speaker: Mr. Wenhao Gao from Department of Chemical Engineering at Massachusetts Institute of Technology (MIT)

We’re super delighted to announce our next speaker, Mr. Wenhao Gao, from Department of Chemical Engineering at Massachusetts Institute of Technology (MIT), to bring us:

“Amortized Tree Generation for Bottom-up Synthesis Planning and Synthesizable Molecular Design”

Abstract

Molecular design and synthesis planning are two critical steps in the process of molecular discovery that we propose to formulate as a single shared task of conditional synthetic pathway generation. We report an amortized approach to generate synthetic pathways as a Markov decision process conditioned on a target molecular embedding. This approach allows us to conduct synthesis planning in a bottom-up manner and design synthesizable molecules by decoding from optimized conditional codes, demonstrating the potential to solve both problems of design and synthesis simultaneously. The approach leverages neural networks to probabilistically model the synthetic trees, one reaction step at a time, according to reactivity rules encoded in a discrete action space of reaction templates. We train these networks on hundreds of thousands of artificial pathways generated from a pool of purchasable compounds and a list of expert-curated templates. We validate our method with (a) the recovery of molecules using conditional generation, (b) the identification of synthesizable structural analogs, and (c) the optimization of molecular structures given oracle functions relevant to drug discovery.

Biography

Wenhao Gao is a Ph.D. candidate in Chemical Engineering at MIT. Under advisory from Prof. Connor Coley, his research focuses on using artificial intelligence and robotic automation to accelerate chemical discovery processes. Before joining the Coley research group, he received his B.S. in Chemistry from Peking University and M.S. in Chemical and Biomolecular Engineering from Johns Hopkins University.

Very much looking forward to the seminar – hope to see you all on Wednesday: (SGT) 9am, Wed, 22 December! https://mit.zoom.us/j/96231985116

Warmest regards, @Thway @kedar @Tonio

Dr. Sergei V. Kalinin (8 December 2021)

Topic: Machine Learning for Scanning Probe and Electron Microscopy for Materials Discovery.

Speaker: Dr. Sergei V. Kalinin from Center for Nanophase Materials Sciences at Oak Ridge National Laboratory

We’re super delighted to announce our next speaker, Dr. Sergei V. Kalinin, from Center for Nanophase Materials Sciences at Oak Ridge National Laboratory to bring us:

“Machine Learning for Scanning Probe and Electron Microscopy for Materials Discovery”

Abstract

Machine learning and artificial intelligence (ML/AI) are rapidly becoming an indispensable part of physics research, with domain applications ranging from theory and materials prediction to high-throughput data analysis. However, the constantly emerging question is how to match the correlative nature of classical ML with hypothesis-driven causal nature of physical sciences. In parallel, the recent successes in applying ML/AI methods for autonomous systems from robotics through self-driving cars to organic and inorganic synthesis are generating enthusiasm for the potential of these techniques to enable automated and autonomous experiment (AE) in imaging.

In this presentation, I will discuss recent progress in automated experiment in electron and scanning probe microscopy, ranging from feature to physics discovery via active learning. The applications of classical deep learning methods in streaming image analysis are strongly affected by the out of distribution drift effects, and the approaches to minimize though are discussed. We further present invariant variational autoencoders as a method to disentangle affine distortions and rotational degrees of freedom from other latent variables in imaging and spectral data. The analysis of the latent space of autoencoders further allows establishing physically relevant transformation mechanisms. Extension of encoder approach towards establishing structure-property relationships will be illustrated on the example of plasmonic structures. I will further illustrate transition from post-experiment data analysis to active learning process, including learning structure-property relationships and materials discovery in composition spread libraries. Here, the strategies based on simple Gaussian Processes often tend to produce sub-optimal results due to the lack of prior knowledge and very simplified (via learned kernel function) representation of spatial complexity of the system. Comparatively, deep kernel learning (DKL) methods allow to realize both the exploration of complex systems towards the discovery of structure-property relationship, and enable automated experiment targeting physics (rather than simple spatial feature) discovery. The latter is illustrated via experimental discovery of the edge plasmons in STEM/EELS and ferroelectric domain dynamics in PFM. For composition spread libraries, I will demonstrate the combination of the structured Gaussian process and reinforcement learning, the approach we refer to as hypothesis learning.

Biography

Sergei Kalinin is a corporate fellow and a group leader at the Center for Nanophase Materials Sciences at Oak Ridge National Laboratory. He received his MS degree from Moscow State University in 1998 and Ph.D. from the University of Pennsylvania (with Dawn Bonnell) in 2002. His research presently focuses on the applications of big data and artificial intelligence methods in atomically resolved imaging by scanning transmission electron microscopy and scanning probes for applications including physics discovery and atomic fabrication, as well as mesoscopic studies of electrochemical, ferroelectric, and transport phenomena via scanning probe microscopy.

Sergei has co-authored >650 publications, with a total citation of >33,000 and an h-index of >94. He is a fellow of MRS, APS, IoP, IEEE, Foresight Institute, and AVS; a recipient of the Blavatnik Award for Physical Sciences (2018), RMS medal for Scanning Probe Microscopy (2015), Presidential Early Career Award for Scientists and Engineers (PECASE) (2009); Burton medal of Microscopy Society of America (2010); 4 R&D100 Awards (2008, 2010, 2016, and 2018); and a number of other distinctions.

Very much looking forward to the seminar – hope to see you all on Wednesday: (SGT) 9am, Wed, 8 December! https://mit.zoom.us/j/96231985116

Warmest regards, @Thway @kedar @Tonio

Mr. Keir Adams (24 November 2021)

Topic: Learning 3D Representations of Molecular Chirality with Invariance to Bond Rotations.

Speaker: Keir Adams from Department of Chemical Engineering and the Center for Computational Science and Engineering at Massachusetts Institute of Technology (MIT)

We’re super delighted to announce our next speaker, Keir Adams, from Department of Chemical Engineering and the Center for Computational Science and Engineering at Massachusetts Institute of Technology (MIT) to bring us:

“Learning 3D Representations of Molecular Chirality with Invariance to Bond Rotations”

Abstract

Molecular stereochemistry strongly alters (bio)chemical interactions yet is often neglected in molecular deep learning. Tetrahedral (point) chirality, a form of stereochemistry describing relative spatial arrangements of bonded neighbors around tetrahedral carbon centers, especially influences substrate-catalyst binding and is therefore critical for pharmaceutical drug design and asymmetric catalysis. However, because chirality influences the set of 3D conformers accessible to the molecule without changing its 2D graph connectivity, modern graph neural networks (GNNs) designed for molecular property prediction often ignore chirality in their learned molecular representations. Most 2D GNNs at best use atomic labels to naïvely treat chirality, while E(3)-invariant 3D GNNs are invariant to chirality altogether.

To enable representation learning on molecules with defined chirality, we design an SE(3)-invariant neural network model that processes torsion angles of a 3D molecular conformer. We explicitly model conformational flexibility by integrating a novel type of invariance to rotations about internal molecular bonds directly into the neural architecture, mitigating the need for multi-conformer data augmentation. We showcase our model on four benchmarks: contrastive learning to distinguish conformers of different stereoisomers in a learned latent space, classification of chiral centers as R/S, prediction of how enantiomers rotate circularly polarized light, and ranking enantiomers by their docking scores in an enantiosensitive protein pocket. We compare our model, Chiral InterRoto Invariant Neural Network (ChIRo), with 2D and 3D GNNs to demonstrate that ChIRo achieves state of the art performance when learning chiral-sensitive functions from molecular structures. We envision that our model, alongside future work to explicitly learn E/Z isomerism and atropisomerism, will expand the capacity of deep learning to advance enantioselective reaction optimization, 3D drug design, and stereochemical property prediction.

Biography

Keir Adams is a 2nd year PhD student in the Department of Chemical Engineering and the Center for Computational Science and Engineering at MIT, advised by Prof. Connor W. Coley. He recently completed his undergraduate degree in chemistry and molecular engineering from the University of Chicago, where he performed research in computational astrochemistry and in renewable energy technologies. At MIT, Keir is generally interested in how 3D molecular representation learning can be used to help accelerate chemicals/materials design and discovery. He is especially interested in developing chemistry-tailored deep learning models that meaningfully understand both molecular stereochemistry and conformer flexibility. In his recent work, “Learning 3D Representations of Molecular Chirality with Invariance to Bond Rotations,” Keir addresses both these themes by designing a neural network that specially encodes 3D torsion angles to meaningfully learn tetrahedral chirality while also introducing a novel invariance to internal bond rotations directly into the model architecture.

Very much looking forward to the seminar – hope to see you all on Wednesday: (SGT) 9am, Wed, 24 November! https://mit.zoom.us/j/96231985116

Warmest regards, @Thway @kedar @Tonio