Crystal structure prediction (CSP) is a useful tool in pharmaceutical development for identifying and assessing risks associated with polymorphism, yet widespread adoption has been hindered by high computational costs and the need for both manual specification and expert knowledge to achieve useful results. Here, we introduce a fully automated, high-throughput CSP protocol designed to overcome these barriers. The protocol's efficiency is driven by Lavo-NN, a novel neural network potential (NNP) architected and trained specifically for pharmaceutical crystal structure generation and ranking. This NNP-driven crystal generation phase is integrated into a scalable cloud-based workflow. We validate this CSP protocol on an extensive retrospective benchmark of 49 unique molecules, almost all of which are drug-like, successfully generating structures that match all 110 $Z' = 1$ experimental polymorphs. The average CSP in this benchmark is performed with approximately 8.4k CPU hours, which is a significant reduction compared to other protocols. The practical utility of the protocol is further demonstrated through case studies that resolve ambiguities in experimental data and a semi-blinded challenge that successfully identifies and ranks polymorphs of three modern drugs from powder X-ray diffraction patterns alone. By significantly reducing the required time and cost, the protocol enables CSP to be routinely deployed earlier in the drug discovery pipeline, such as during lead optimization. Rapid turnaround times and high throughput also enable CSP that can be run in parallel with experimental screening, providing chemists with real-time insights to guide their work in the lab.

We extend BeamAttack, an adversarial attack algorithm designed to evaluate the robustness of text classification systems through word-level modifications guided by beam search. Our extensions include support for word deletions and the option to skip substitutions, enabling the discovery of minimal modifications that alter model predictions. We also integrate LIME to better prioritize word replacements. Evaluated across multiple datasets and victim models (BiLSTM, BERT, and adversarially trained RoBERTa) within the BODEGA framework, our approach achieves over a 99\% attack success rate while preserving the semantic and lexical similarity of the original texts. Through both quantitative and qualitative analysis, we highlight BeamAttack's effectiveness and its limitations. Our implementation is available at https://github.com/LucK1Y/BeamAttack

Neural network potentials (NNPs) offer a powerful alternative to traditional force fields for molecular dynamics (MD) simulations. Accurate and stable MD simulations, crucial for evaluating material properties, require training data encompassing both low-energy stable structures and high-energy structures. Conventional knowledge distillation (KD) methods fine-tune a pre-trained NNP as a teacher model to generate training data for a student model. However, in material-specific models, this fine-tuning process increases energy barriers, making it difficult to create training data containing high-energy structures. To address this, we propose a novel KD framework that leverages a non-fine-tuned, off-the-shelf pre-trained NNP as a teacher. Its gentler energy landscape facilitates the exploration of a wider range of structures, including the high-energy structures crucial for stable MD simulations. Our framework employs a two-stage training process: first, the student NNP is trained with a dataset generated by the off-the-shelf teacher; then, it is fine-tuned with a smaller, high-accuracy density functional theory (DFT) dataset. We demonstrate the effectiveness of our framework by applying it to both organic (polyethylene glycol) and inorganic (L$_{10}$GeP$_{2}$S$_{12}$) materials, achieving comparable or superior accuracy in reproducing physical properties compared to existing methods. Importantly, our method reduces the number of expensive DFT calculations by 10x compared to existing NNP generation methods, without sacrificing accuracy. Furthermore, the resulting student NNP achieves up to 106x speedup in inference compared to the teacher NNP, enabling significantly faster and more efficient MD simulations.

Neural network potentials (NNPs) offer a fast and accurate alternative to ab-initio methods for molecular dynamics (MD) simulations but are hindered by the high cost of training data from high-fidelity Quantum Mechanics (QM) methods. Our work introduces the Implicit Delta Learning (IDLe) method, which reduces the need for high-fidelity QM data by leveraging cheaper semi-empirical QM computations without compromising NNP accuracy or inference cost. IDLe employs an end-to-end multi-task architecture with fidelity-specific heads that decode energies based on a shared latent representation of the input atomistic system. In various settings, IDLe achieves the same accuracy as single high-fidelity baselines while using up to 50x less high-fidelity data. This result could significantly reduce data generation cost and consequently enhance accuracy and generalization, and expand chemical coverage for NNPs, advancing MD simulations for material science and drug discovery. Additionally, we provide a novel set of 11 million semi-empirical QM calculations to support future multi-fidelity NNP modeling.

Machine learning potentials offer a revolutionary, unifying framework for molecular simulations across scales, from quantum chemistry to coarse-grained models. Here, I explore their potential to dramatically improve accuracy and scalability in simulating complex molecular systems. I discuss key challenges that must be addressed to fully realize their transformative potential in chemical biology and related fields. Machine learning (ML) potentials are poised to revolutionize molecular simulations across multiple scales, leveraging the innate ability of neural networks to capture complex correlations in high-dimensional spaces. Here, I discuss how these advanced models can dramatically improve the accuracy and efficiency of simulations, from quantum-mechanical calculations to coarse-grained dynamics. By bridging the gap between atomistic detail and macroscopic behavior, ML potentials promise to unlock new insights into molecular processes, drug discovery, and materials design [Duignan24]. I highlight recent successes, current challenges, and future directions in this rapidly evolving field, emphasizing the transformative potential of ML-driven simulations in chemical biology and related disciplines.

Machine learning (ML) techniques and atomistic modeling have rapidly transformed materials design and discovery. Specifically, generative models can swiftly propose promising materials for targeted applications. However, the predicted properties of materials through the generative models often do not match with calculated properties through ab initio calculations. This discrepancy can arise because the generated coordinates are not fully relaxed, whereas the many properties are derived from relaxed structures. Neural network-based potentials (NNPs) can expedite the process by providing relaxed structures from the initially generated ones. Nevertheless, acquiring data to train NNPs for this purpose can be extremely challenging as it needs to encompass previously unknown structures. This study utilized extended ensemble molecular dynamics (MD) to secure a broad range of liquid- and solid-phase configurations in one of the metallic systems, nickel. Then, we could significantly reduce them through active learning without losing much accuracy. We found that the NNP trained from the distilled data could predict different energy-minimized closed-pack crystal structures even though those structures were not explicitly part of the initial data. Furthermore, the data can be translated to other metallic systems (aluminum and niobium), without repeating the sampling and distillation processes. Our approach to data acquisition and distillation has demonstrated the potential to expedite NNP development and enhance materials design and discovery by integrating generative models.

Developing accurate and efficient coarse-grained representations of proteins is crucial for understanding their folding, function, and interactions over extended timescales. Our methodology involves simulating proteins with molecular dynamics and utilizing the resulting trajectories to train a neural network potential through differentiable trajectory reweighting. Remarkably, this method requires only the native conformation of proteins, eliminating the need for labeled data derived from extensive simulations or memory-intensive end-to-end differentiable simulations. Once trained, the model can be employed to run parallel molecular dynamics simulations and sample folding events for proteins both within and beyond the training distribution, showcasing its extrapolation capabilities. By applying Markov State Models, native-like conformations of the simulated proteins can be predicted from the coarse-grained simulations. Owing to its theoretical transferability and ability to use solely experimental static structures as training data, we anticipate that this approach will prove advantageous for developing new protein force fields and further advancing the study of protein dynamics, folding, and interactions.

Traditional sequence-to-sequence (seq2seq) models and other variations of the attention-mechanism such as hierarchical attention have been applied to the text summarization problem. Though there is a hierarchy in the way humans use language by forming paragraphs from sentences and sentences from words, hierarchical models have usually not worked that much better than their traditional seq2seq counterparts. This effect is mainly because either the hierarchical attention mechanisms are too sparse using hard attention or noisy using soft attention. In this paper, we propose a method based on extracting the highlights of a document; a key concept that is conveyed in a few sentences. In a typical text summarization dataset consisting of documents that are 800 tokens in length (average), capturing long-term dependencies is very important, e.g., the last sentence can be grouped with the first sentence of a document to form a summary. LSTMs (Long Short-Term Memory) proved useful for machine translation. However, they often fail to capture long-term dependencies while modeling long sequences. To address these issues, we have adapted Neural Semantic Encoders (NSE) to text summarization, a class of memory-augmented neural networks by improving its functionalities and proposed a novel hierarchical NSE that outperforms similar previous models significantly. The quality of summarization was improved by augmenting linguistic factors, namely lemma, and Part-of-Speech (PoS) tags, to each word in the dataset for improved vocabulary coverage and generalization. The hierarchical NSE model on factored dataset outperformed the state-of-the-art by nearly 4 ROUGE points. We further designed and used the first GPU-based self-critical Reinforcement Learning model.

Highly accurate potential energy surfaces are of key interest for the detailed understanding and predictive modeling of chemical systems. In recent years, several new types of force fields, which are based on machine learning algorithms and fitted to ab initio reference calculations, have been introduced to meet this requirement. Here we show how high-dimensional neural network potentials can be employed to automatically generate the potential energy surface of finite sized clusters at coupled cluster accuracy, namely CCSD(T*)-F12a/aug-cc-pVTZ. The developed automated procedure utilizes the established intrinsic properties of the model such that the configurations for the training set are selected in an unbiased and efficient way to minimize the computational effort of expensive reference calculations. These ideas are applied to protonated water clusters from the hydronium cation, H$_3$O$^+$, up to the tetramer, H$_9$O$_{4}^{+}$, and lead to a common potential energy surface that describes all these systems at essentially converged coupled cluster accuracy with a fitting error of 0.06 kJ/mol per atom. The fit is validated in detail and separately for all clusters up to the tetramer and yields reliable results not only for stationary points, but also for reaction pathways, intermediate configurations, as well as different sampling techniques. Per design the NNPs constructed in this fashion can handle very different conditions including the quantum nature of the nuclei and enhanced sampling techniques covering very low as well as high temperatures. This enables fast and exhaustive exploration of the targeted protonated water clusters at essentially converged interactions. In addition, the automated process will allow one to tackle finite systems much beyond the present case.

Molecular dynamics simulations are an important tool for describing the evolution of a chemical system with time. However, these simulations are inherently held back either by the prohibitive cost of accurate electronic structure theory computations or the limited accuracy of classical empirical force fields. Machine learning techniques can help to overcome these limitations by providing access to potential energies, forces and other molecular properties modeled directly after an electronic structure reference at only a fraction of the original computational cost. The present text discusses several practical aspects of conducting machine learning driven molecular dynamics simulations. First, we study the efficient selection of reference data points on the basis of an active learning inspired adaptive sampling scheme. This is followed by the analysis of a machine-learning based model for simulating molecular dipole moments in the framework of predicting infrared spectra via molecular dynamics simulations. Finally, we show that machine learning models can offer valuable aid in understanding chemical systems beyond a simple prediction of quantities.