Clinical Decision Support Systems (CDSS) play an increasingly important role in medical diagnostics. We present AI Diagnostic Assistant (AIDA), a real-time predictive model designed to assist doctors in interpreting patient conditions. AIDA analyzes electronic health records (EHR), including medical history, laboratory results, and complaints, to suggest potential diagnoses from 95 common conditions before the doctor makes the final decision. The model acts as a verification and backup tool, ensuring that no critical details are overlooked. Trained on 1.5 million patient records and validated on a dataset curated by a panel of experts, AIDA proves trustworthy as a diagnosis-making assistant (87.7% accuracy compared to 91.7% accuracy among doctors). Integrated into a megapolis-wide CDSS, AIDA has assisted doctors in over 3 million real-world diagnoses to date.

Age accelerating condition Healthy control p-value Age acceleration Blood data C h r o n o lo g ic a l A g e CpG site Dataset Averaging across cells ... ... ... ... ... CpG N CpG ...Figure 1: ComputAgeBench: the first comprehensive framework for benchmarking various epigenetic aging clock models.For a dataset 𝑋 , obtained by profiling DNA methylation at CpG sites in bulk blood samples, an aging clock model 𝑓 is trained to distinguish healthy individuals from those with pre-defined aging-accelerating conditions.

Adaptive deep brain stimulation (aDBS) has emerged as a promising treatment for Parkinson's disease (PD).In aDBS, a surgically placed electrode sends dynamically altered stimuli to the brain based on neurophysiological feedback: an invasive gadget that limits the amount of data one could collect for optimizing the control offline.As a consequence, a plethora of synthetic models of PD and those of the control algorithms have been proposed.Herein, we introduce the first neurophysiologically realistic benchmark for comparing said models.Specifically, our methodology covers not only conventional basal ganglia circuit dynamics and pathological oscillations, but also captures 15 previously dismissed physiological attributes, such as signal instabilities and noise, neural drift, electrode conductance changes and individual variability -all modeled as spatially distributed and temporally registered features via beta-band activity in the brain and a feedback.Furthermore, we purposely built our framework as a structured environment for training and evaluating deep reinforcement learning (RL) algorithms, opening new possibilities for optimizing aDBS control strategies and inviting the machine learning community to contribute to the emerging field of intelligent neurostimulation interfaces.Code repository

We present a new approach for Neural Optimal Transport (NOT) training procedure, capable of accurately and efficiently estimating optimal transportation plan via specific regularization on dual Kantorovich potentials. The main bottleneck of existing NOT solvers is associated with the procedure of finding a near-exact approximation of the conjugate operator (i.e., the c-transform), which is done either by optimizing over non-convex max-min objectives or by the computationally intensive fine-tuning of the initial approximated prediction. We resolve both issues by proposing a new theoretically justified loss in the form of expectile regularization which enforces binding conditions on the learning process of the dual potentials. Such a regularization provides the upper bound estimation over the distribution of possible conjugate potentials and makes the learning stable, completely eliminating the need for additional extensive fine-tuning. Proposed method, called Expectile-Regularized Neural Optimal Transport (ENOT), outperforms previous state-of-the-art approaches in the established Wasserstein-2 benchmark tasks by a large margin (up to a 3-fold improvement in quality and up to a 10-fold improvement in runtime). Moreover, we showcase performance of ENOT for various cost functions in different tasks, such as image generation, demonstrating generalizability and robustness of the proposed algorithm.

This paper derives a new strong Gaussian approximation bound for the sum of independent random vectors. The approach relies on the optimal transport theory and yields explicit dependence on the dimension size p and the sample size n. This dependence establishes a new fundamental limit for all practical applications of statistical learning theory. Particularly, based on this bound, we prove approximation in distribution for the maximum norm in a high-dimensional setting (p > n).

Image corruption by motion artifacts is an ingrained problem in Magnetic Resonance Imaging (MRI). In this work, we propose a neural network-based regularization term to enhance Autofocusing, a classic optimization-based method to remove motion artifacts. The method takes the best of both worlds: the optimization-based routine iteratively executes the blind demotion and deep learning-based prior penalizes for unrealistic restorations and speeds up the convergence. We validate the method on three models of motion trajectories, using synthetic and real noisy data. The method proves resilient to noise and anatomic structure variation, outperforming the state-of-the-art demotion methods.

Detecting Out-of-Distribution (OoD) data is one of the greatest challenges in safe and robust deployment of machine learning algorithms in medicine. When the algorithms encounter cases that deviate from the distribution of the training data, they often produce incorrect and over-confident predictions. OoD detection algorithms aim to catch erroneous predictions in advance by analysing the data distribution and detecting potential instances of failure. Moreover, flagging OoD cases may support human readers in identifying incidental findings. Due to the increased interest in OoD algorithms, benchmarks for different domains have recently been established. In the medical imaging domain, for which reliable predictions are often essential, an open benchmark has been missing. We introduce the Medical-Out-Of-Distribution-Analysis-Challenge (MOOD) as an open, fair, and unbiased benchmark for OoD methods in the medical imaging domain. The analysis of the submitted algorithms shows that performance has a strong positive correlation with the perceived difficulty, and that all algorithms show a high variance for different anomalies, making it yet hard to recommend them for clinical practice. We also see a strong correlation between challenge ranking and performance on a simple toy test set, indicating that this might be a valuable addition as a proxy dataset during anomaly detection algorithm development.

Accurate interpretation of images of complex joints, such as the temporomandibular joint (TMJ), has become essential in a variety of clinical practices, ranging from the basic assessment of wear and tear (e.g., osteoarthritis) to intricate surgical interventions (e.g., arthroplasty). Today, this examination remains subjective and time-consuming, requiring comprehensive understanding of the joint’s properties. Ultrasound (US) is the main medical imaging modality that, in addition to its many advantages, allows assessing the condition of the joint during its physiological movement. Therefore, there is a demand for an automatic and efficient method to track the joint landmarks and their movements in the ultrasound videos, promising facilitation of the tedious routine work of the US operators and a more objective metric of the joint health. To address the problem of landmark detection, we propose a method that combines 3D U-Net (which extracts spatial patterns with abstract features) and Long Short-Term Memory module (LSTM, for processing data as temporal 2D sequences of video frames). The method is evaluated on the dataset containing 13 sequences of TMJ motion recordings during the opening and the closing movements of the lower jaw. The approach proved functional and could be readily integrated into the current clinical practice.

Artem Shelmanov, Dmitri Puzyrev, Lyubov Kupriyanova, Denis Belyakov, Daniil Larionov, Nikita Khromov, Olga Kozlova, Ekaterina Artemova, Dmitry V. Dylov, Alexander Panchenko. Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume. 2021.

Peripheral Difficult Venous Access (PDVA) is a commonplace problem in clinical practice which results in repetitive punctures, damaged veins, and significant discomfort to the patients. Nowadays, the poor visibility of subcutaneous vasculature in the visible part of the light spectrum is overcome by near-infrared (NIR) imaging and a returned projection of the recognized vasculature back to the arm of the patient. We introduce the first “smart” engine to govern the components of such imagers in a mixed reality setting. Namely, a closed-loop hardware system that optimizes cross-talk between the virtual mask generated from the NIR measurement and the projected augmenting image is proposed. Such real-virtual image translation is accomplished in several steps. First, the NIR vein segmentation task is solved using U-Net-based network architecture and the Frangi vesselness filter. The generated mask is then transformed and translated into the visible domain by a projector that adjusts for distortions and misalignment with the true vasculature using the paradigm of Reinforcement Learning (RL). We propose a new class of mixed reality reward functions that guarantees proper alignment of the projected image regardless of angle, translation, and scale offsets between the NIR measurement and the visible projection.

Malfunctioning neurons in the brain sometimes operate synchronously, reportedly causing many neurological diseases, e.g. Parkinson’s. Suppression and control of this collective synchronous activity are therefore of great importance for neuroscience, and can only rely on limited engineering trials due to the need to experiment with live human brains. We present the first Reinforcement Learning (RL) gym framework that emulates this collective behavior of neurons and allows us to find suppression parameters for the environment of synthetic degenerate models of neurons. We successfully suppress synchrony via RL for three pathological signaling regimes, characterize the framework’s stability to noise, and further remove the unwanted oscillations by engaging multiple PPO agents.

We propose a new unsupervised and non-parametric method to detect change points in electrocardiography. The detection relies on optimal transport theory combined with topological analysis and the bootstrap procedure. The algorithm is designed to detect changes in virtually any harmonic or a partially harmonic signal and is verified on ECG data streams. We successfully find abnormal or irregular cardiac cycles in the waveforms for the six of the most frequent types of clinical arrhythmias using a single algorithm. Our unsupervised approach reaches the level of performance of the supervised state-of-the-art techniques. We provide conceptual justification for the efficiency of the method.

Approaches to constructing statistical phantoms (SPs) for the needs of data fusion with the real data have attracted a lot of attention in the recent years. SPs assist with radiation dosimetry and provide metrics for estimating image quality in the medical X-ray systems. Unlike existing SPs, we propose to consider both material electron densities and anatomical size variations models to portray realistic variations in radiologic data. We introduce a new adaptive unsupervised fusion approach for data augmentation, capable of generating a variety of medically adequate images. The model is based on a combination of continuous Poisson modelling of voxel values, Monte Carlo rejection sampling scheme, and a landmark-based warping. Unlike mere average of HU values of each organ (typical in the other state-of-art SPs), our augmented voxels depict intensity fluctuations, effectively mimicking a distribution of electron densities within each organ. In the experimental section, we evaluate the proposed method and demonstrate its superiority compared to the existing methods. This phantom generation could be instrumental for assessing dose uncertainty, unsupervised refinement of image reconstruction, image classification, and semantic segmentation tasks carried out by machine learning algorithms in the scenarios of limited available data.

The quality and interpretability of the state-of-the-art methods for automatic analysis of chest X-ray images is still not sufficient. We address this problem by presenting a model that combines the analysis of frontal chest X-ray scans with structured patient information contained within radiology records. The proposed model generates a short textual summary with essential information on the found pathologies along with their location and severity; and the 2D heatmaps localizing each pathology on the original X-ray images. We test the proposed model on the MIMIC-CXR dataset. It achieves the state-of-the-art performance for image labelling and captioning (78.5% of correctly generated sentences) and defeats other similar solutions that dismiss the additional patient data (by 5.2% of correctly generated sentences). We also propose an automatic approach to label mining that leverages multimodal data: the X-ray images, related textual reports, patients' age and sex.

Active learning is a technique that helps to minimize the annotation budget required for the creation of a labeled dataset while maximizing the performance of a model trained on this dataset. It has been shown that active learning can be successfully applied to sequence tagging tasks of text processing in conjunction with deep learning models even when a limited amount of labeled data is available. Recent advances in transfer learning methods for natural language processing based on deep pre-trained models such as ELMo and BERT offer a much better ability to generalize on small annotated datasets compared to their shallow counterparts. The combination of deep pre-trained models and active learning leads to a powerful approach to dealing with annotation scarcity. In this work, we investigate the potential of this approach on clinical and biomedical data. The experimental evaluation shows that the combination of active learning and deep pre-trained models outperforms the standard methods of active learning. We also suggest a modification to a standard uncertainty sampling strategy and empirically show that it could be beneficial for annotation of very skewed datasets. Finally, we propose an annotation tool empowered with active learning and deep pre-trained models that could be used for entity annotation directly from Jupyter IDE.