Search

Predicting protein functional characteristics from structure remains a central problem in protein science, with broad implications from understanding the mechanisms of disease to designing novel therapeutics. Unfortunately, current machine learning methods are limited by scarce and biased experimental data, and physics-based methods are either too slow to be useful, or too simplified to be accurate. In this work, we present Loop-Diffusion, an energy based diffusion model which leverages a dataset of general protein loops from the entire protein universe to learn an energy function that generalizes to functional prediction tasks. We evaluate Loop-Diffusion's performance on scoring TCR-pMHC interfaces and demonstrate state-of-the-art results in recognizing binding-enhancing mutations.

The explainability of machine learning algorithms is crucial, and numerous methods have emerged recently. Local, post-hoc methods assign an attribution score to each feature, indicating its importance for the prediction. However, these methods require recalculating explanations for each example. On the other side, while there exist global approaches they often produce explanations that are either overly simplistic and unreliable or excessively complex. To bridge this gap, we propose GLEAMS, a novel method that partitions the input space and learns an interpretable model within each sub-region, thereby providing both faithful local and global surrogates. We demonstrate GLEAMS' effectiveness on both synthetic and real-world data, highlighting its desirable properties and human-understandable insights.

Predicting the stability and fitness effects of amino acid mutations in proteins is a cornerstone of biological discovery and engineering. Various experimental techniques have been developed to measure mutational effects, providing us with extensive datasets across a diverse range of proteins. By training on these data, traditional computational modeling and more recent machine learning approaches have advanced significantly in predicting mutational effects. Here, we introduce HERMES, a 3D rotationally equivariant structure-based neural network model for mutational effect and stability prediction. Pre-trained to predict amino acid propensity from its surrounding 3D structure, HERMES can be fine-tuned for mutational effects using our open-source code. We present a suite of HERMES models, pre-trained with different strategies, and fine-tuned to predict the stability effect of mutations. Benchmarking against other models shows that HERMES often outperforms or matches their performance in predicting mutational effect on stability, binding, and fitness. HERMES offers versatile tools for evaluating mutational effects and can be fine-tuned for specific predictive objectives., Comment: 16 pages, 8 figures

Accurately modeling protein 3D structure is essential for the design of functional proteins. An important sub-task of structure modeling is protein side-chain packing: predicting the conformation of side-chains (rotamers) given the protein's backbone structure and amino-acid sequence. Conventional approaches for this task rely on expensive sampling procedures over hand-crafted energy functions and rotamer libraries. Recently, several deep learning methods have been developed to tackle the problem in a data-driven way, albeit with vastly different formulations (from image-to-image translation to directly predicting atomic coordinates). Here, we frame the problem as a joint regression over the side-chains' true degrees of freedom: the dihedral $\chi$ angles. We carefully study possible objective functions for this task, while accounting for the underlying symmetries of the task. We propose Holographic Packer (H-Packer), a novel two-stage algorithm for side-chain packing built on top of two light-weight rotationally equivariant neural networks. We evaluate our method on CASP13 and CASP14 targets. H-Packer is computationally efficient and shows favorable performance against conventional physics-based algorithms and is competitive against alternative deep learning solutions., Comment: Accepted as a conference paper at MLCB 2023. 8 pages main body, 20 pages with appendix. 10 figures

Recent advances in the development of ultra-low loss silicon nitride integrated photonic circuits have heralded a new generation of integrated lasers capable of reaching fiber laser coherence. However, these devices presently are based on self-injection locking of distributed feedback (DFB) laser diodes, increasing both the cost and requiring tuning of laser setpoints for their operation. In contrast, turn-key legacy laser systems use reflective semiconductor optical amplifiers (RSOA). While this scheme has been utilized for integrated photonics-based lasers, so far, no cost-effective RSOA-based integrated lasers exist that are low noise and simultaneously feature fast, mode-hop-free and linear frequency tuning as required for frequency modulated continuous wave (FMCW) LiDAR or for laser locking in frequency metrology. Here we overcome this challenge and demonstrate a RSOA-based, frequency agile integrated laser, that can be tuned with high speed, with high linearity at low power. This is achieved using monolithic integration of piezoelectrical actuators on ultra-low loss silicon nitride photonic integrated circuits in a Vernier filter-based laser scheme. The laser operates at 1550 nm, features 6 mW output power, 400 Hz intrinsic laser linewidth, and allows ultrafast wavelength switching within 7 ns rise time and 75 nW power consumption. In addition, we demonstrate the suitability for FMCW LiDAR by showing laser frequency tuning over 1.5 GHz at 100 kHz triangular chirp rate with nonlinearity of 0.25% after linearization, and use the source for measuring a target scene 10 m away with a 8.5 cm distance resolution.

Group-equivariant neural networks have emerged as a data-efficient approach to solve classification and regression tasks, while respecting the relevant symmetries of the data. However, little work has been done to extend this paradigm to the unsupervised and generative domains. Here, we present Holographic-(Variational) Auto Encoder (H-(V)AE), a fully end-to-end SO(3)-equivariant (variational) autoencoder in Fourier space, suitable for unsupervised learning and generation of data distributed around a specified origin in 3D. H-(V)AE is trained to reconstruct the spherical Fourier encoding of data, learning in the process a low-dimensional representation of the data (i.e., a latent space) with a maximally informative rotationally invariant embedding alongside an equivariant frame describing the orientation of the data. We extensively test the performance of H-(V)AE on diverse datasets. We show that the learned latent space efficiently encodes the categorical features of spherical images. Moreover, H-(V)AE's latent space can be used to extract compact embeddings for protein structure microenvironments, and when paired with a Random Forest Regressor, it enables state-of-the-art predictions of protein-ligand binding affinity.

The switch from a Model-Centric to a Data-Centric mindset is putting emphasis on data and its quality rather than algorithms, bringing forward new challenges. In particular, the sensitive nature of the information in highly regulated scenarios needs to be accounted for. Specific approaches to address the privacy issue have been developed, as Privacy Enhancing Technologies. However, they frequently cause loss of information, putting forward a crucial trade-off among data quality and privacy. A clever way to bypass such a conundrum relies on Synthetic Data: data obtained from a generative process, learning the real data properties. Both Academia and Industry realized the importance of evaluating synthetic data quality: without all-round reliable metrics, the innovative data generation task has no proper objective function to maximize. Despite that, the topic remains under-explored. For this reason, we systematically catalog the important traits of synthetic data quality and privacy, and devise a specific methodology to test them. The result is DAISYnt (aDoption of Artificial Intelligence SYnthesis): a comprehensive suite of advanced tests, which sets a de facto standard for synthetic data evaluation. As a practical use-case, a variety of generative algorithms have been trained on real-world Credit Bureau Data. The best model has been assessed, using DAISYnt on the different synthetic replicas. Further potential uses, among others, entail auditing and fine-tuning of generative models or ensuring high quality of a given synthetic dataset. From a prescriptive viewpoint, eventually, DAISYnt may pave the way to synthetic data adoption in highly regulated domains, ranging from Finance to Healthcare, through Insurance and Education.

Face Presentation Attack Detection (PAD) is an important measure to prevent spoof attacks for face biometric systems. Many works based on Convolution Neural Networks (CNNs) for face PAD formulate the problem as an image-level binary classification task without considering the context. Alternatively, Vision Transformers (ViT) using self-attention to attend the context of an image become the mainstreams in face PAD. Inspired by ViT, we propose a Video-based Transformer for face PAD (ViTransPAD) with short/long-range spatio-temporal attention which can not only focus on local details with short attention within a frame but also capture long-range dependencies over frames. Instead of using coarse image patches with single-scale as in ViT, we propose the Multi-scale Multi-Head Self-Attention (MsMHSA) architecture to accommodate multi-scale patch partitions of Q, K, V feature maps to the heads of transformer in a coarse-to-fine manner, which enables to learn a fine-grained representation to perform pixel-level discrimination for face PAD. Due to lack inductive biases of convolutions in pure transformers, we also introduce convolutions to the proposed ViTransPAD to integrate the desirable properties of CNNs by using convolution patch embedding and convolution projection. The extensive experiments show the effectiveness of our proposed ViTransPAD with a preferable accuracy-computation balance, which can serve as a new backbone for face PAD.

In strained mechanical resonators, the concurrence of tensile stress and geometric nonlinearity dramatically reduces dissipation. This phenomenon, dissipation dilution, is employed in mirror suspensions of gravitational wave interferometers and at the nanoscale, where soft-clamping and strain engineering have allowed extremely high quality factors. However, these techniques have so far only been applied in amorphous materials, specifically silicon nitride. Crystalline materials exhibit significantly lower intrinsic damping at cryogenic temperatures, due to the absence of two level systems in the bulk, as exploited in Weber bars and silicon optomechanical cavities. Applying dissipation dilution engineering to strained crystalline materials could therefore enable extremely low loss nanomechanical resonators, due to the combination of reduced internal friction, high intrinsic strain, and high yield strength. Pioneering work has not yet fully exploited this potential. Here, we demonstrate that single crystal strained silicon, a material developed for high mobility transistors, can be used to realize mechanical resonators with ultralow dissipation. We observe that high aspect ratio ($>10^5$) strained silicon nanostrings support MHz mechanical modes with quality factors exceeding $10^{10}$ at 7 K, a tenfold improvement over values reported in silicon nitride. At 7 K, the thermal noise-limited force sensitivity is approximately $45\ \mathrm{{zN}/{\sqrt{Hz}}}$ - approaching that of carbon nanotubes - and the heating rate is only 60 quanta-per-second. Our nanomechanical resonators exhibit lower dissipation than the most pristine macroscopic oscillators and their low mass makes them particularly promising for quantum sensing and transduction., Comment: 15 pages, 11 figures

We address the problem of modeling constrained hospital resources in the midst of the COVID-19 pandemic in order to inform decision-makers of future demand and assess the societal value of possible interventions. For broad applicability, we focus on the common yet challenging scenario where patient-level data for a region of interest are not available. Instead, given daily admissions counts, we model aggregated counts of observed resource use, such as the number of patients in the general ward, in the intensive care unit, or on a ventilator. In order to explain how individual patient trajectories produce these counts, we propose an aggregate count explicit-duration hidden Markov model, nicknamed the ACED-HMM, with an interpretable, compact parameterization. We develop an Approximate Bayesian Computation approach that draws samples from the posterior distribution over the model's transition and duration parameters given aggregate counts from a specific location, thus adapting the model to a region or individual hospital site of interest. Samples from this posterior can then be used to produce future forecasts of any counts of interest. Using data from the United States and the United Kingdom, we show our mechanistic approach provides competitive probabilistic forecasts for the future even as the dynamics of the pandemic shift. Furthermore, we show how our model provides insight about recovery probabilities or length of stay distributions, and we suggest its potential to answer challenging what-if questions about the societal value of possible interventions., Comment: To appear in the Proceedings of the Machine Learning for Healthcare (MLHC) conference, 2021. 20 pages, 7 figures and 1 table. 26 additional pages of supplementary material

In the global economy, credit companies play a central role in economic development, through their activity as money lenders. This important task comes with some drawbacks, mainly the risk of the debtors not being able to repay the provided credit. Therefore, Credit Risk Modelling (CRM), namely the evaluation of the probability that a debtor will not repay the due amount, plays a paramount role. Statistical approaches have been successfully exploited since long, becoming the most used methods for CRM. Recently, also machine and deep learning techniques have been applied to the CRM task, showing an important increase in prediction quality and performances. However, such techniques usually do not provide reliable explanations for the scores they come up with. As a consequence, many machine and deep learning techniques fail to comply with western countries' regulations such as, for example, GDPR. In this paper we suggest to use LIME (Local Interpretable Model-agnostic Explanations) technique to tackle the explainability problem in this field, we show its employment on a real credit-risk dataset and eventually discuss its soundness and the necessary improvements to guarantee its adoption and compliance with the task.

The aim of this project is to develop and test advanced analytical methods to improve the prediction accuracy of Credit Risk Models, preserving at the same time the model interpretability. In particular, the project focuses on applying an explainable machine learning model to bank-related databases. The input data were obtained from open data. Over the total proven models, CatBoost has shown the highest performance. The algorithm implementation produces a GINI of 0.68 after tuning the hyper-parameters. SHAP package is used to provide a global and local interpretation of the model predictions to formulate a human-comprehensive approach to understanding the decision-maker algorithm. The 20 most important features are selected using the Shapley values to present a full human-understandable model that reveals how the attributes of an individual are related to its model prediction.

To study neurodegenerative diseases, longitudinal studies are carried on volunteer patients. During a time span of several months to several years, they go through regular medical visits to acquire data from different modalities, such as biological samples, cognitive tests, structural and functional imaging. These variables are heterogeneous but they all depend on the patient's health condition, meaning that there are possibly unknown relationships between all modalities. Some information may be specific to some modalities, others may be complementary, and others may be redundant. Some data may also be missing. In this work we present a neural network architecture for multimodal learning, able to use imaging and clinical data from two time points to predict the evolution of a neurodegenerative disease, and robust to missing values. Our multimodal network achieves 92.5\% accuracy and an AUC score of 0.978 over a test set of 57 subjects. We also show the superiority of the multimodal architecture, for up to 37.5\% of missing values in test set subjects' clinical measurements, compared to a model using only the clinical modality., Comment: Accepted in the 10th International Conference on Image Processing Theory, Tools and Applications (IPTA 2020)