Partitioning variability in animal behavioral videos using semi-supervised variational autoencoders

Recent neuroscience studies demonstrate that a deeper understanding of brain function requires a deeper understanding of behavior. Detailed behavioral measurements are now often collected using video cameras, resulting in an increased need for computer vision algorithms that extract useful information from video data. Here we introduce a new video analysis tool that combines the output of supervised pose estimation algorithms (e.g. DeepLabCut) with unsupervised dimensionality reduction methods to produce interpretable, low-dimensional representations of behavioral videos that extract more information than pose estimates alone. We demonstrate this tool by extracting interpretable behavioral features from videos of three different head-fixed mouse preparations, as well as a freely moving mouse in an open field arena, and show how these interpretable features can facilitate downstream behavioral and neural analyses. We also show how the behavioral features produced by our model improve the precision and interpretation of these downstream analyses compared to using the outputs of either fully supervised or fully unsupervised methods alone.

Predicting likelihood of psychological disorders in PlayerUnknown's Battlegrounds (PUBG) players from Asian countries using supervised machine learning

Multiplayer Online Battle Arena (MOBA) has become one of the most popular genre of online video games played by gamers worldwide. Previous studies have exhibited that excessive engagement in games can lead to Internet Gaming Disorder (IGD). Internet Gaming Disorder has been associated with psychological disorders like impulsivity, anxiety and Attention Deficit Hyperactivity Disorder (ADHD). In this study, we propose an approach to use the game and player statistics along with self-esteem measure of a PlayerUnknown's Battlegrounds (PUBG, a MOBA game) player to predict whether he/she suffers from IGD and psychological disorders namely ADHD and Generalized Anxiety Disorder (GAD). We extract the game and player statistics of PUBG players from Asian countries and then run several state of the art supervised machine learning models to predict the occurrence of IGD, ADHD, and GAD. Initial experiments and results show that we are able to predict IGD, ADHD, and GAD with an accuracy of 93.18%, 81.81% and 84.9% respectively. Game statistics of PUBG players show strong positive correlation with IGD and ADHD indicating detrimental effects of MOBA games.

Streaming chunk incremental learning for class-wise data stream classification with fast learning speed and low structural complexity

Due to the fast speed of data generation and collection from advanced equipment, the amount of data obviously overflows the limit of available memory space and causes difficulties achieving high learning accuracy. Several methods based on discard-after-learn concept have been proposed. Some methods were designed to cope with a single incoming datum but some were designed for a chunk of incoming data. Although the results of these approaches are rather impressive, most of them are based on temporally adding more neurons to learn new incoming data without any neuron merging process which can obviously increase the computational time and space complexities. Only online versatile elliptic basis function (VEBF) introduced neuron merging to reduce the space-time complexity of learning only a single incoming datum. This paper proposed a method for further enhancing the capability of discard-after-learn concept for streaming data-chunk environment in terms of low computational time and neural space complexities. A set of recursive functions for computing the relevant parameters of a new neuron, based on statistical confidence interval, was introduced. The newly proposed method, named streaming chunk incremental learning (SCIL), increases the plasticity and the adaptabilty of the network structure according to the distribution of incoming data and their classes. When being compared to the others in incremental-like manner, based on 11 benchmarked data sets of 150 to 581,012 samples with attributes ranging from 4 to 1,558 formed as streaming data, the proposed SCIL gave better accuracy and time in most data sets.

Rapid detection of internalizing diagnosis in young children enabled by wearable sensors and machine learning

There is a critical need for fast, inexpensive, objective, and accurate screening tools for childhood psychopathology. Perhaps most compelling is in the case of internalizing disorders, like anxiety and depression, where unobservable symptoms cause children to go unassessed-suffering in silence because they never exhibiting the disruptive behaviors that would lead to a referral for diagnostic assessment. If left untreated these disorders are associated with long-term negative outcomes including substance abuse and increased risk for suicide. This paper presents a new approach for identifying children with internalizing disorders using an instrumented 90-second mood induction task. Participant motion during the task is monitored using a commercially available wearable sensor. We show that machine learning can be used to differentiate children with an internalizing diagnosis from controls with 81% accuracy (67% sensitivity, 88% specificity). We provide a detailed description of the modeling methodology used to arrive at these results and explore further the predictive ability of each temporal phase of the mood induction task. Kinematical measures most discriminative of internalizing diagnosis are analyzed in detail, showing affected children exhibit significantly more avoidance of ambiguous threat. Performance of the proposed approach is compared to clinical thresholds on parent-reported child symptoms which differentiate children with an internalizing diagnosis from controls with slightly lower accuracy (.68-.75 vs. .81), slightly higher specificity (.88-1.00 vs. .88), and lower sensitivity (.00-.42 vs. .67) than the proposed, instrumented method. These results point toward the future use of this approach for screening children for internalizing disorders so that interventions can be deployed when they have the highest chance for long-term success.

PLATYPUS: A Multiple-View Learning Predictive Framework for Cancer Drug Sensitivity Prediction

Cancer is a complex collection of diseases that are to some degree unique to each patient. Precision oncology aims to identify the best drug treatment regime using molecular data on tumor samples. While omics-level data is becoming more widely available for tumor specimens, the datasets upon which computational learning methods can be trained vary in coverage from sample to sample and from data type to data type. Methods that can 'connect the dots' to leverage more of the information provided by these studies could offer major advantages for maximizing predictive potential. We introduce a multi-view machinelearning strategy called PLATYPUS that builds 'views' from multiple data sources that are all used as features for predicting patient outcomes. We show that a learning strategy that finds agreement across the views on unlabeled data increases the performance of the learning methods over any single view. We illustrate the power of the approach by deriving signatures for drug sensitivity in a large cancer cell line database. Code and additional information are available from the PLATYPUS website https://sysbiowiki.soe.ucsc.edu/platypus.

A supervised joint multi-layer segmentation framework for retinal optical coherence tomography images using conditional random field

BACKGROUND AND OBJECTIVE

Accurate segmentation of the intra-retinal tissue layers in Optical Coherence Tomography (OCT) images plays an important role in the diagnosis and treatment of ocular diseases such as Age-Related Macular Degeneration (AMD) and Diabetic Macular Edema (DME). The existing energy minimization based methods employ multiple, manually handcrafted cost terms and often fail in the presence of pathologies. In this work, we eliminate the need to handcraft the energy by learning it from training images in an end-to-end manner. Our method can be easily adapted to pathologies by re-training it on an appropriate dataset.

METHODS

We propose a Conditional Random Field (CRF) framework for the joint multi-layer segmentation of OCT B-scans. The appearance of each retinal layer and boundary is modeled by two convolutional filter banks and the shape priors are modeled using Gaussian distributions. The total CRF energy is linearly parameterized to allow a joint, end-to-end training by employing the Structured Support Vector Machine formulation.

RESULTS

The proposed method outperformed three benchmark algorithms on four public datasets. The NORMAL-1 and NORMAL-2 datasets contain healthy OCT B-scans while the AMD-1 and DME-1 dataset contain B-scans of AMD and DME cases respectively. The proposed method achieved an average unsigned boundary localization error (U-BLE) of 1.52 pixels on NORMAL-1, 1.11 pixels on NORMAL-2 and 2.04 pixels on the combined NORMAL-1 and DME-1 dataset across the eight layer boundaries, outperforming the three benchmark methods in each case. The Dice coefficient was 0.87 on NORMAL-1, 0.89 on NORMAL-2 and 0.84 on the combined NORMAL-1 and DME-1 dataset across the seven retinal layers. On the combined NORMAL-1 and AMD-1 dataset, we achieved an average U-BLE of 1.86 pixels on the ILM, inner and outer RPE boundaries and a Dice of 0.98 for the ILM-RPE_in region and 0.81 for the RPE layer.

CONCLUSION

We have proposed a supervised CRF based method to jointly segment multiple tissue layers in OCT images. It can aid the ophthalmologists in the quantitative analysis of structural changes in the retinal tissue layers for clinical practice and large-scale clinical studies.

Faceness-Net: Face Detection through Deep Facial Part Responses

We propose a deep convolutional neural network (CNN) for face detection leveraging on facial attributes based supervision. We observe a phenomenon that part detectors emerge within CNN trained to classify attributes from uncropped face images, without any explicit part supervision. The observation motivates a new method for finding faces through scoring facial parts responses by their spatial structure and arrangement. The scoring mechanism is data-driven, and carefully formulated considering challenging cases where faces are only partially visible. This consideration allows our network to detect faces under severe occlusion and unconstrained pose variations. Our method achieves promising performance on popular benchmarks including FDDB, PASCAL Faces, AFW, and WIDER FACE.

Discriminatively Trained Latent Ordinal Model for Video Classification

We address the problem of video classification for facial analysis and human action recognition. We propose a novel weakly supervised learning method that models the video as a sequence of automatically mined, discriminative sub-events (e.g., onset and offset phase for "smile", running and jumping for "highjump"). The proposed model is inspired by the recent works on Multiple Instance Learning and latent SVM/HCRF - it extends such frameworks to model the ordinal aspect in the videos, approximately. We obtain consistent improvements over relevant competitive baselines on four challenging and publicly available video based facial analysis datasets for prediction of expression, clinical pain and intent in dyadic conversations, and on three challenging human action datasets. We also validate the method with qualitative results and show that they largely support the intuitions behind the method.

Supervised machine learning reveals introgressed loci in the genomes of Drosophila simulans and D. sechellia

Hybridization and gene flow between species appears to be common. Even though it is clear that hybridization is widespread across all surveyed taxonomic groups, the magnitude and consequences of introgression are still largely unknown. Thus it is crucial to develop the statistical machinery required to uncover which genomic regions have recently acquired haplotypes via introgression from a sister population. We developed a novel machine learning framework, called FILET (Finding Introgressed Loci via Extra-Trees) capable of revealing genomic introgression with far greater power than competing methods. FILET works by combining information from a number of population genetic summary statistics, including several new statistics that we introduce, that capture patterns of variation across two populations. We show that FILET is able to identify loci that have experienced gene flow between related species with high accuracy, and in most situations can correctly infer which population was the donor and which was the recipient. Here we describe a data set of outbred diploid Drosophila sechellia genomes, and combine them with data from D. simulans to examine recent introgression between these species using FILET. Although we find that these populations may have split more recently than previously appreciated, FILET confirms that there has indeed been appreciable recent introgression (some of which might have been adaptive) between these species, and reveals that this gene flow is primarily in the direction of D. simulans to D. sechellia.

Understanding the extent to which species or diverged populations hybridize in nature is crucially important if we are to understand the speciation process. Accordingly numerous research groups have developed methodology for finding the genetic evidence of such introgression. In this report we develop a supervised machine learning approach for uncovering loci which have introgressed across species boundaries. We show that our method, FILET, has greater accuracy and power than competing methods in discovering introgression, and in addition can detect the directionality associated with the gene flow between species. Using whole genome sequences from Drosophila simulans and Drosophila sechellia we show that FILET discovers quite extensive introgression between these species that has occurred mostly from D. simulans to D. sechellia. Our work highlights the complex process of speciation even within a well-studied system and points to the growing importance of supervised machine learning in population genetics.

Mapping Patient Trajectories using Longitudinal Extraction and Deep Learning in the MIMIC-III Critical Care Database

Electronic Health Records (EHRs) contain a wealth of patient data useful to biomedical researchers. At present, both the extraction of data and methods for analyses are frequently designed to work with a single snapshot of a patient's record. Health care providers often perform and record actions in small batches over time. By extracting these care events, a sequence can be formed providing a trajectory for a patient's interactions with the health care system. These care events also offer a basic heuristic for the level of attention a patient receives from health care providers. We show that is possible to learn meaningful embeddings from these care events using two deep learning techniques, unsupervised autoencoders and long short-term memory networks. We compare these methods to traditional machine learning methods which require a point in time snapshot to be extracted from an EHR.

Data-driven advice for applying machine learning to bioinformatics problems

As the bioinformatics field grows, it must keep pace not only with new data but with new algorithms. Here we contribute a thorough analysis of 13 state-of-the-art, commonly used machine learning algorithms on a set of 165 publicly available classification problems in order to provide data-driven algorithm recommendations to current researchers. We present a number of statistical and visual comparisons of algorithm performance and quantify the effect of model selection and algorithm tuning for each algorithm and dataset. The analysis culminates in the recommendation of five algorithms with hyperparameters that maximize classifier performance across the tested problems, as well as general guidelines for applying machine learning to supervised classification problems.

Considerations for automated machine learning in clinical metabolic profiling: Altered homocysteine plasma concentration associated with metformin exposure

With the maturation of metabolomics science and proliferation of biobanks, clinical metabolic profiling is an increasingly opportunistic frontier for advancing translational clinical research. Automated Machine Learning (AutoML) approaches provide exciting opportunity to guide feature selection in agnostic metabolic profiling endeavors, where potentially thousands of independent data points must be evaluated. In previous research, AutoML using high-dimensional data of varying types has been demonstrably robust, outperforming traditional approaches. However, considerations for application in clinical metabolic profiling remain to be evaluated. Particularly, regarding the robustness of AutoML to identify and adjust for common clinical confounders. In this study, we present a focused case study regarding AutoML considerations for using the Tree-Based Optimization Tool (TPOT) in metabolic profiling of exposure to metformin in a biobank cohort. First, we propose a tandem rank-accuracy measure to guide agnostic feature selection and corresponding threshold determination in clinical metabolic profiling endeavors. Second, while AutoML, using default parameters, demonstrated potential to lack sensitivity to low-effect confounding clinical covariates, we demonstrated residual training and adjustment of metabolite features as an easily applicable approach to ensure AutoML adjustment for potential confounding characteristics. Finally, we present increased homocysteine with long-term exposure to metformin as a potentially novel, non-replicated metabolite association suggested by TPOT; an association not identified in parallel clinical metabolic profiling endeavors. While warranting independent replication, our tandem rank-accuracy measure suggests homocysteine to be the metabolite feature with largest effect, and corresponding priority for further translational clinical research. Residual training and adjustment for a potential confounding effect by BMI only slightly modified the suggested association. Increased homocysteine is thought to be associated with vitamin B12 deficiency - evaluation for potential clinical relevance is suggested. While considerations for clinical metabolic profiling are recommended, including adjustment approaches for clinical confounders, AutoML presents an exciting tool to enhance clinical metabolic profiling and advance translational research endeavors.

Improving the explainability of Random Forest classifier - user centered approach

Machine Learning (ML) methods are now influencing major decisions about patient care, new medical methods, drug development and their use and importance are rapidly increasing in all areas. However, these ML methods are inherently complex and often difficult to understand and explain resulting in barriers to their adoption and validation. Our work (RFEX) focuses on enhancing Random Forest (RF) classifier explainability by developing easy to interpret explainability summary reports from trained RF classifiers as a way to improve the explainability for (often non-expert) users. RFEX is implemented and extensively tested on Stanford FEATURE data where RF is tasked with predicting functional sites in 3D molecules based on their electrochemical signatures (features). In developing RFEX method we apply user-centered approach driven by explainability questions and requirements collected by discussions with interested practitioners. We performed formal usability testing with 13 expert and non-expert users to verify RFEX usefulness. Analysis of RFEX explainability report and user feedback indicates its usefulness in significantly increasing explainability and user confidence in RF classification on FEATURE data. Notably, RFEX summary reports easily reveal that one needs very few (from 2-6 depending on a model) top ranked features to achieve 90% or better of the accuracy when all 480 features are used.

Toward more realistic drug-target interaction predictions

A number of supervised machine learning models have recently been introduced for the prediction of drug-target interactions based on chemical structure and genomic sequence information. Although these models could offer improved means for many network pharmacology applications, such as repositioning of drugs for new therapeutic uses, the prediction models are often being constructed and evaluated under overly simplified settings that do not reflect the real-life problem in practical applications. Using quantitative drug-target bioactivity assays for kinase inhibitors, as well as a popular benchmarking data set of binary drug-target interactions for enzyme, ion channel, nuclear receptor and G protein-coupled receptor targets, we illustrate here the effects of four factors that may lead to dramatic differences in the prediction results: (i) problem formulation (standard binary classification or more realistic regression formulation), (ii) evaluation data set (drug and target families in the application use case), (iii) evaluation procedure (simple or nested cross-validation) and (iv) experimental setting (whether training and test sets share common drugs and targets, only drugs or targets or neither). Each of these factors should be taken into consideration to avoid reporting overoptimistic drug-target interaction prediction results. We also suggest guidelines on how to make the supervised drug-target interaction prediction studies more realistic in terms of such model formulations and evaluation setups that better address the inherent complexity of the prediction task in the practical applications, as well as novel benchmarking data sets that capture the continuous nature of the drug-target interactions for kinase inhibitors.