Changes over time in brain anatomy can provide important insight for treatment design or scientific analyses. We present a method that predicts how a brain MRI for an individual will change over time. We model changes using a diffeomorphic deformation field that we predict using function using convolutional neural networks. Given a predicted deformation field, a baseline scan can be warped to give a prediction of the brain scan at a future time. We demonstrate the method using the ADNI cohort, and analyze how performance is affected by model variants and the subject-specific information provided. We show that the model provides good predictions and that external clinical data can improve predictions.
Animals have been a common sighting on roads in India which leads to several accidents between them and vehicles every year. This makes it vital to develop a support system for driverless vehicles that assists in preventing these forms of accidents. In this paper, we propose a neoteric framework for avoiding vehicle-to-animal collisions by developing an efficient approach for the detection of animals on highways using deep learning and computer vision techniques on dashcam video. Our approach leverages the Mask R-CNN model for detecting and identifying various commonly found animals. Then, we perform lane detection to deduce whether a detected animal is on the vehicle's lane or not and track its location and direction of movement using a centroid based object tracking algorithm. This approach ensures that the framework is effective at determining whether an animal is obstructing the path or not of an autonomous vehicle in addition to predicting its movement and giving feedback accordingly. This system was tested under various lighting and weather conditions and was observed to perform relatively well, which leads the way for prominent driverless vehicle's support systems for avoiding vehicular collisions with animals on Indian roads in real-time.
Recent advances in trajectory replanning have enabled quadrotor to navigate autonomously in unknown environments. However, high-speed navigation still remains a significant challenge. Given very limited time, existing methods have no strong guarantee on the feasibility or quality of the solutions. Moreover, most methods do not consider environment perception, which is the key bottleneck to fast flight. In this paper, we present RAPTOR, a robust and perception-aware replanning framework to support fast and safe flight. A path-guided optimization (PGO) approach that incorporates multiple topological paths is devised, to ensure finding feasible and high-quality trajectories in very limited time. We also introduce a perception-aware planning strategy to actively observe and avoid unknown obstacles. A risk-aware trajectory refinement ensures that unknown obstacles which may endanger the quadrotor can be observed earlier and avoid in time. The motion of yaw angle is planned to actively explore the surrounding space that is relevant for safe navigation. The proposed methods are tested extensively. We will release our implementation as an open-source package for the community.
To date, the privacy-protection intended pixelation tasks are still labor-intensive and yet to be studied. With the prevailing of video live streaming, establishing an online face pixelation mechanism during streaming is an urgency. In this paper, we develop a new method called Face Pixelation in Video Live Streaming (FPVLS) to generate automatic personal privacy filtering during unconstrained streaming activities. Simply applying multi-face trackers will encounter problems in target drifting, computing efficiency, and over-pixelation. Therefore, for fast and accurate pixelation of irrelevant people's faces, FPVLS is organized in a frame-to-video structure of two core stages. On individual frames, FPVLS utilizes image-based face detection and embedding networks to yield face vectors. In the raw trajectories generation stage, the proposed Positioned Incremental Affinity Propagation (PIAP) clustering algorithm leverages face vectors and positioned information to quickly associate the same person's faces across frames. Such frame-wise accumulated raw trajectories are likely to be intermittent and unreliable on video level. Hence, we further introduce the trajectory refinement stage that merges a proposal network with the two-sample test based on the Empirical Likelihood Ratio (ELR) statistic to refine the raw trajectories. A Gaussian filter is laid on the refined trajectories for final pixelation. On the video live streaming dataset we collected, FPVLS obtains satisfying accuracy, real-time efficiency, and contains the over-pixelation problems.
We propose a self-supervised approach for learning representations and robotic behaviors entirely from unlabeled videos recorded from multiple viewpoints, and study how this representation can be used in two robotic imitation settings: imitating object interactions from videos of humans, and imitating human poses. Imitation of human behavior requires a viewpoint-invariant representation that captures the relationships between end-effectors (hands or robot grippers) and the environment, object attributes, and body pose. We train our representations using a metric learning loss, where multiple simultaneous viewpoints of the same observation are attracted in the embedding space, while being repelled from temporal neighbors which are often visually similar but functionally different. In other words, the model simultaneously learns to recognize what is common between different-looking images, and what is different between similar-looking images. This signal causes our model to discover attributes that do not change across viewpoint, but do change across time, while ignoring nuisance variables such as occlusions, motion blur, lighting and background. We demonstrate that this representation can be used by a robot to directly mimic human poses without an explicit correspondence, and that it can be used as a reward function within a reinforcement learning algorithm. While representations are learned from an unlabeled collection of task-related videos, robot behaviors such as pouring are learned by watching a single 3rd-person demonstration by a human. Reward functions obtained by following the human demonstrations under the learned representation enable efficient reinforcement learning that is practical for real-world robotic systems. Video results, open-source code and dataset are available at https://sermanet.github.io/imitate
Machine learning techniques applied to the Natural Language Processing (NLP) component of conversational agent development show promising results for improved accuracy and quality of feedback that a conversational agent can provide. The effort required to develop an educational scenario specific conversational agent is time consuming as it requires domain experts to label and annotate noisy data sources such as classroom videos. Previous approaches to modeling annotations have relied on labeling thousands of examples and calculating inter-annotator agreement and majority votes in order to model the necessary scenarios. This method, while proven successful, ignores individual annotator strengths in labeling a data point and under-utilizes examples that do not have a majority vote for labeling. We propose using a multi-task weak supervision method combined with active learning to address these concerns. This approach requires less labeling than traditional methods and shows significant improvements in precision, efficiency, and time-requirements than the majority vote method (Ratner 2019). We demonstrate the validity of this method on the Google Jigsaw data set and then propose a scenario to apply this method using the Instructional Quality Assessment(IQA) to define the categories for labeling. We propose using probabilistic modeling of annotator labeling to generate active learning examples to further label the data. Active learning is able to iteratively improve the training performance and accuracy of the original classification model. This approach combines state-of-the art labeling techniques of weak supervision and active learning to optimize results in the educational domain and could be further used to lessen the data requirements for expanded scenarios within the education domain through transfer learning.
Variation in speech is often represented and investigated using phonetic transcriptions, but transcribing speech is time-consuming and error prone. To create reliable representations of speech independent from phonetic transcriptions, we investigate the extraction of acoustic embeddings from several self-supervised neural models. We use these representations to compute word-based pronunciation differences between non-native and native speakers of English, and evaluate these differences by comparing them with human native-likeness judgments. We show that Transformer-based speech representations lead to significant performance gains over the use of phonetic transcriptions, and find that feature-based use of Transformer models is most effective with one or more middle layers instead of the final layer. We also demonstrate that these neural speech representations not only capture segmental differences, but also intonational and durational differences that cannot be represented by a set of discrete symbols used in phonetic transcriptions.
Identification of a linear time-invariant dynamical system from partial observations is a fundamental problem in control theory. A natural question is how to do so with non-asymptotic statistical rates depending on the inherent dimensionality (order) $d$ of the system, rather than on the sufficient rollout length or on $\frac1{1-\rho(A)}$, where $\rho(A)$ is the spectral radius of the dynamics matrix. We develop the first algorithm that given a single trajectory of length $T$ with gaussian observation noise, achieves a near-optimal rate of $\widetilde O\left(\sqrt\frac{d}{T}\right)$ in $\mathcal{H}_2$ error for the learned system. We also give bounds under process noise and improved bounds for learning a realization of the system. Our algorithm is based on low-rank approximation of Hankel matrices of geometrically increasing sizes.
Detecting substructure within strongly lensed images is a promising route to shed light on the nature of dark matter. It is a challenging task, which traditionally requires detailed lens modeling and source reconstruction, taking weeks to analyze each system. We use machine learning to circumvent the need for lens and source modeling and develop a method to both locate subhalos in an image as well as determine their mass using the technique of image segmentation. The network is trained on images with a single subhalo located near the Einstein ring. Training in this way allows the network to learn the gravitational lensing of light and it is then able to accurately detect entire populations of substructure, even far from the Einstein ring. In images with a single subhalo and without noise, the network detects subhalos of mass $10^6 M_{\odot}$ 62% of the time and 78% of these detected subhalos are predicted in the correct mass bin. The detection accuracy increases for heavier masses. When random noise at the level of 1% of the mean brightness of the image is included (which is a realistic approximation HST, for sources brighter than magnitude 20), the network loses sensitivity to the low-mass subhalos; with noise, the $10^{8.5}M_{\odot}$ subhalos are detected 86% of the time, but the $10^8 M_{\odot}$ subhalos are only detected 38% of the time. The false-positive rate is around 2 false subhalos per 100 images with and without noise, coming mostly from masses $\leq10^8 M_{\odot}$. With good accuracy and a low false-positive rate, counting the number of pixels assigned to each subhalo class over multiple images allows for a measurement of the subhalo mass function (SMF). When measured over five mass bins from $10^8 M_{\odot}$ to $10^{10} M_{\odot}$ the SMF slope is recovered with an error of 14.2 (16.3)% for 10 images, and this improves to 2.1 (2.6)% for 1000 images without (with 1%) noise.
Achieving the full promise of the Thermodynamic Variational Objective (TVO), a recently proposed variational lower bound on the log evidence involving a one-dimensional Riemann integral approximation, requires choosing a "schedule" of sorted discretization points. This paper introduces a bespoke Gaussian process bandit optimization method for automatically choosing these points. Our approach not only automates their one-time selection, but also dynamically adapts their positions over the course of optimization, leading to improved model learning and inference. We provide theoretical guarantees that our bandit optimization converges to the regret-minimizing choice of integration points. Empirical validation of our algorithm is provided in terms of improved learning and inference in Variational Autoencoders and Sigmoid Belief Networks.