Creative Technologies (United States)

Stan is a probabilistic programming language for specifying statistical models. A Stan program imperatively defines a log probability function over parameters conditioned on specified data and constants. As of version 2.14.0, Stan provides full Bayesian inference for continuous-variable models through Markov chain Monte Carlo methods such as the No-U-Turn sampler, an adaptive form of Hamiltonian Monte Carlo sampling. Penalized maximum likelihood estimates are calculated using optimization methods such as the limited memory Broyden-Fletcher-Goldfarb-Shanno algorithm. Stan is also a platform for computing log densities and their gradients and Hessians, which can be used in alternative algorithms such as variational Bayes, expectation propagation, and marginal inference using approximate integration. To this end, Stan is set up so that the densities, gradients, and Hessians, along with intermediate quantities of the algorithm such as acceptance probabilities, are easily accessible. Stan can be called from the command line using the cmdstan package, through R using the rstan package, and through Python using the pystan package. All three interfaces support sampling and optimization-based inference with diagnostics and posterior analysis. rstan and pystan also provide access to log probabilities, gradients, Hessians, parameter transforms, and specialized plotting.

This article is a survey of methods for measuring agreement among corpus annotators. It exposes the mathematics and underlying assumptions of agreement coefficients, covering Krippendorff's alpha as well as Scott's pi and Cohen's kappa; discusses the use of coefficients in several annotation tasks; and argues that weighted, alpha-like coefficients, traditionally less used than kappa-like measures in computational linguistics, may be more appropriate for many corpus annotation tasks—but that their use makes the interpretation of the value of the coefficient even harder.

Intelligent control may be viewed as the ability of a controller to operate in multiple environments by recognizing which environment is currently in existence and servicing it appropriately. An important prerequisite for an intelligent controller is the ability to adapt rapidly to any unknown but constant operating environment. This paper presents a general methodology for such adaptive control using multiple models, switching, and tuning. The approach was first introduced by Narendra et al. (1992) for improving the transient response of adaptive systems in a stable fashion. This paper proposes different switching and tuning schemes for adaptive control which combine fixed and adaptive models in novel ways. The principal mathematical results are the proofs of stability when these different schemes are used in the context of model reference control of an unknown linear time-invariant system. A variety of simulation results are presented to demonstrate the efficacy of the proposed methods.

We introduce Pixel-aligned Implicit Function (PIFu), an implicit representation that locally aligns pixels of 2D images with the global context of their corresponding 3D object. Using PIFu, we propose an end-to-end deep learning method for digitizing highly detailed clothed humans that can infer both 3D surface and texture from a single image, and optionally, multiple input images. Highly intricate shapes, such as hairstyles, clothing, as well as their variations and deformations can be digitized in a unified way. Compared to existing representations used for 3D deep learning, PIFu produces high-resolution surfaces including largely unseen regions such as the back of a person. In particular, it is memory efficient unlike the voxel representation, can handle arbitrary topology, and the resulting surface is spatially aligned with the input image. Furthermore, while previous techniques are designed to process either a single image or multiple views, PIFu extends naturally to arbitrary number of views. We demonstrate high-resolution and robust reconstructions on real world images from the DeepFashion dataset, which contains a variety of challenging clothing types. Our method achieves state-of-the-art performance on a public benchmark and outperforms the prior work for clothed human digitization from a single image.

The field of 3D face modeling has a large gap between high-end and low-end methods. At the high end, the best facial animation is indistinguishable from real humans, but this comes at the cost of extensive manual labor. At the low end, face capture from consumer depth sensors relies on 3D face models that are not expressive enough to capture the variability in natural facial shape and expression. We seek a middle ground by learning a facial model from thousands of accurately aligned 3D scans. Our FLAME model (Faces Learned with an Articulated Model and Expressions) is designed to work with existing graphics software and be easy to fit to data. FLAME uses a linear shape space trained from 3800 scans of human heads. FLAME combines this linear shape space with an articulated jaw, neck, and eyeballs, pose-dependent corrective blendshapes, and additional global expression blendshapes. The pose and expression dependent articulations are learned from 4D face sequences in the D3DFACS dataset along with additional 4D sequences. We accurately register a template mesh to the scan sequences and make the D3DFACS registrations available for research purposes. In total the model is trained from over 33, 000 scans. FLAME is low-dimensional but more expressive than the FaceWarehouse model and the Basel Face Model. We compare FLAME to these models by fitting them to static 3D scans and 4D sequences using the same optimization method. FLAME is significantly more accurate and is available for research purposes (http://flame.is.tue.mpg.de).

In neural networks, it is often desirable to work with various representations of the same space. For example, 3D rotations can be represented with quaternions or Euler angles. In this paper, we advance a definition of a continuous representation, which can be helpful for training deep neural networks. We relate this to topological concepts such as homeomorphism and embedding. We then investigate what are continuous and discontinuous representations for 2D, 3D, and n-dimensional rotations. We demonstrate that for 3D rotations, all representations are discontinuous in the real Euclidean spaces of four or fewer dimensions. Thus, widely used representations such as quaternions and Euler angles are discontinuous and difficult for neural networks to learn. We show that the 3D rotations have continuous representations in 5D and 6D, which are more suitable for learning. We also present continuous representations for the general case of the n-dimensional rotation group SO(n). While our main focus is on rotations, we also show that our constructions apply to other groups such as the orthogonal group and similarity transforms. We finally present empirical results, which show that our continuous rotation representations outperform discontinuous ones for several practical problems in graphics and vision, including a simple autoencoder sanity test, a rotation estimator for 3D point clouds, and an inverse kinematics solver for 3D human poses.

Recent advances in deep learning have shown exciting promise in filling large holes in natural images with semantically plausible and context aware details, impacting fundamental image manipulation tasks such as object removal. While these learning-based methods are significantly more effective in capturing high-level features than prior techniques, they can only handle very low-resolution inputs due to memory limitations and difficulty in training. Even for slightly larger images, the inpainted regions would appear blurry and unpleasant boundaries become visible. We propose a multi-scale neural patch synthesis approach based on joint optimization of image content and texture constraints, which not only preserves contextual structures but also produces high-frequency details by matching and adapting patches with the most similar mid-layer feature correlations of a deep classification network. We evaluate our method on the ImageNet and Paris Streetview datasets and achieved state-of-the-art inpainting accuracy. We show our approach produces sharper and more coherent results than prior methods, especially for high-resolution images.

We introduce a method to render Neural Radiance Fields (NeRFs) in real time using PlenOctrees, an octree-based 3D representation which supports view-dependent effects. Our method can render 800×800 images at more than 150 FPS, which is over 3000 times faster than conventional NeRFs. We do so without sacrificing quality while preserving the ability of NeRFs to perform free-viewpoint rendering of scenes with arbitrary geometry and view-dependent effects. Real-time performance is achieved by pre-tabulating the NeRF into a PlenOctree. In order to preserve view-dependent effects such as specularities, we factorize the appearance via closed-form spherical basis functions. Specifically, we show that it is possible to train NeRFs to predict a spherical harmonic representation of radiance, removing the viewing direction as an input to the neural network. Furthermore, we show that PlenOctrees can be directly optimized to further minimize the reconstruction loss, which leads to equal or better quality compared to competing methods. Moreover, this octree optimization step can be used to reduce the training time, as we no longer need to wait for the NeRF training to converge fully. Our real-time neural rendering approach may potentially enable new applications such as 6-DOF industrial and product visualizations, as well as next generation AR/VR systems. PlenOctrees are amenable to in-browser rendering as well; please visit the project page for the interactive online demo, as well as video and code: https://alexyu.net/plenoctrees.

The use of head-mounted displays (HMD) for virtual reality (VR) application-based purposes including therapy, rehabilitation, and training is increasing. Despite advancements in VR technologies, many users still experience sickness symptoms. VR sickness may be influenced by technological differences within HMDs such as resolution and refresh rate, however, VR content also plays a significant role. The primary objective of this systematic review and meta-analysis was to examine the literature on HMDs that report Simulator Sickness Questionnaire (SSQ) scores to determine the impact of content. User factors associated with VR sickness were also examined. A systematic search was conducted according to PRISMA guidelines. Fifty-five articles met inclusion criteria, representing 3,016 participants (mean age range 19.5-80; 41% female). Findings show gaming content recorded the highest total SSQ mean 34.26 (95%CI 29.57-38.95). VR sickness profiles were also influenced by visual stimulation, locomotion and exposure times. Older samples (mean age ≥35 years) scored significantly lower total SSQ means than younger samples, however, these findings are based on a small evidence base as a limited number of studies included older users. No sex differences were found. Across all types of content, the pooled total SSQ mean was relatively high 28.00 (95%CI 24.66-31.35) compared with recommended SSQ cut-off scores. These findings are of relevance for informing future research and the application of VR in different contexts.

Speech processing algorithms are often developed demonstrating improvements over the state-of-the-art, but sometimes at the cost of high complexity. This makes algorithm reimplementations based on literature difficult, and thus reliable comparisons between published results and current work are hard to achieve. This paper presents a new collaborative and freely available repository for speech processing algorithms called COVAREP, which aims at fast and easy access to new speech processing algorithms and thus facilitating research in the field. We envisage that COVAREP will allow more reproducible research by strengthening complex implementations through shared contributions and openly available code which can be discussed, commented on and corrected by the community. Presently COVAREP contains contributions from five distinct laboratories and we encourage contributions from across the speech processing research field. In this paper, we provide an overview of the current offerings of COVAREP and also include a demonstration of the algorithms through an emotion classification experiment.

Rendering bridges the gap between 2D vision and 3D scenes by simulating the physical process of image formation. By inverting such renderer, one can think of a learning approach to infer 3D information from 2D images. However, standard graphics renderers involve a fundamental discretization step called rasterization, which prevents the rendering process to be differentiable, hence able to be learned. Unlike the state-of-the-art differentiable renderers, which only approximate the rendering gradient in the back propagation, we propose a truly differentiable rendering framework that is able to (1) directly render colorized mesh using differentiable functions and (2) back-propagate efficient supervision signals to mesh vertices and their attributes from various forms of image representations, including silhouette, shading and color images. The key to our framework is a novel formulation that views rendering as an aggregation function that fuses the probabilistic contributions of all mesh triangles with respect to the rendered pixels. Such formulation enables our framework to flow gradients to the occluded and far-range vertices, which cannot be achieved by the previous state-of-the-arts. We show that by using the proposed renderer, one can achieve significant improvement in 3D unsupervised single-view reconstruction both qualitatively and quantitatively. Experiments also demonstrate that our approach is able to handle the challenging tasks in image-based shape fitting, which remain nontrivial to existing differentiable renderers. Code is available at https://github.com/ShichenLiu/SoftRas.

We present SimSensei Kiosk, an implemented virtual human interviewer designed to create an engaging face-to-face inter-action where the user feels comfortable talking and sharing information. SimSensei Kiosk is also designed to create in-teractional situations favorable to the automatic assessment of distress indicators, defined as verbal and nonverbal behav-iors correlated with depression, anxiety or post-traumatic stress disorder (PTSD). In this paper, we summarize the de-sign methodology, performed over the past two years, which is based on three main development cycles: (1) analysis of face-to-face human interactions to identify potential distress indicators, dialogue policies and virtual human gestures, (2) development and analysis of a Wizard-of-Oz prototype sys-tem where two human operators were deciding the spoken and gestural responses, and (3) development of a fully au-tomatic virtual interviewer able to engage users in 15-25 minute interactions. We show the potential of our fully auto-matic virtual human interviewer in a user study, and situate its performance in relation to the Wizard-of-Oz prototype.

The Audio/Visual Emotion Challenge and Workshop (AVEC 2016) "Depression, Mood and Emotion" will be the sixth competition event aimed at comparison of multimedia processing and machine learning methods for automatic audio, visual and physiological depression and emotion analysis, with all participants competing under strictly the same conditions. The goal of the Challenge is to provide a common benchmark test set for multi-modal information processing and to bring together the depression and emotion recognition communities, as well as the audio, video and physiological processing communities, to compare the relative merits of the various approaches to depression and emotion recognition under well-defined and strictly comparable conditions and establish to what extent fusion of the approaches is possible and beneficial. This paper presents the challenge guidelines, the common data used, and the performance of the baseline system on the two tasks.

We introduce an architecture and toolkit for building dialogue managers currently being developed in the TRINDI project, based on the notions of information state and dialogue move engine. The aim is to provide a framework for experimenting with implementations of different theories of information state, information state update and dialogue control. A number of dialogue managers are currently being built using the toolkit, and we present overviews of two of them. We believe that this framework will make implementation of dialogue processing theories easier, also facilitating comparison of different types of dialogue systems, thus helping to achieve a prerequisite for arriving at a best practice for the development of the dialogue management component of a spoken dialogue system.

With more than 10,000 new videos posted online every day on social websites such as YouTube and Facebook, the internet is becoming an almost infinite source of information. One crucial challenge for the coming decade is to be able to harvest relevant information from this constant flow of multimodal data. This paper addresses the task of multimodal sentiment analysis, and conducts proof-of-concept experiments that demonstrate that a joint model that integrates visual, audio, and textual features can be effectively used to identify sentiment in Web videos. This paper makes three important contributions. First, it addresses for the first time the task of tri-modal sentiment analysis, and shows that it is a feasible task that can benefit from the joint exploitation of visual, audio and textual modalities. Second, it identifies a subset of audio-visual features relevant to sentiment analysis and present guidelines on how to integrate these features. Finally, it introduces a new dataset consisting of real online data, which will be useful for future research in this area.

Facial feature detection algorithms have seen great progress over the recent years. However, they still struggle in poor lighting conditions and in the presence of extreme pose or occlusions. We present the Constrained Local Neural Field model for facial landmark detection. Our model includes two main novelties. First, we introduce a probabilistic patch expert (landmark detector) that can learn non-linear and spatial relationships between the input pixels and the probability of a landmark being aligned. Secondly, our model is optimised using a novel Non-uniform Regularised Landmark Mean-Shift optimisation technique, which takes into account the reliabilities of each patch expert. We demonstrate the benefit of our approach on a number of publicly available datasets over other state-of-the-art approaches when performing landmark detection in unseen lighting conditions and in the wild.

The use of the commercial video games as rehabilitation tools, such as the Nintendo WiiFit, has recently gained much interest in the physical therapy arena. Motion tracking controllers such as the Nintendo Wiimote are not sensitive enough to accurately measure performance in all components of balance. Additionally, users can figure out how to "cheat" inaccurate trackers by performing minimal movement (e.g. wrist twisting a Wiimote instead of a full arm swing). Physical rehabilitation requires accurate and appropriate tracking and feedback of performance. To this end, we are developing applications that leverage recent advances in commercial video game technology to provide full-body control of animated virtual characters. A key component of our approach is the use of newly available low cost depth sensing camera technology that provides markerless full-body tracking on a conventional PC. The aim of this research was to develop and assess an interactive game-based rehabilitation tool for balance training of adults with neurological injury.

BACKGROUND: Therapeutic virtual reality (VR) has emerged as an efficacious treatment modality for a wide range of health conditions. However, despite encouraging outcomes from early stage research, a consensus for the best way to develop and evaluate VR treatments within a scientific framework is needed. OBJECTIVE: We aimed to develop a methodological framework with input from an international working group in order to guide the design, implementation, analysis, interpretation, and communication of trials that develop and test VR treatments. METHODS: A group of 21 international experts was recruited based on their contributions to the VR literature. The resulting Virtual Reality Clinical Outcomes Research Experts held iterative meetings to seek consensus on best practices for the development and testing of VR treatments. RESULTS: The interactions were transcribed, and key themes were identified to develop a scientific framework in order to support best practices in methodology of clinical VR trials. Using the Food and Drug Administration Phase I-III pharmacotherapy model as guidance, a framework emerged to support three phases of VR clinical study designs-VR1, VR2, and VR3. VR1 studies focus on content development by working with patients and providers through the principles of human-centered design. VR2 trials conduct early testing with a focus on feasibility, acceptability, tolerability, and initial clinical efficacy. VR3 trials are randomized, controlled studies that evaluate efficacy against a control condition. Best practice recommendations for each trial were provided. CONCLUSIONS: Patients, providers, payers, and regulators should consider this best practice framework when assessing the validity of VR treatments.

Discusses some of the key issues that must be addressed in creating virtual humans, or androids. As a first step, we overview the issues and available tools in three key areas of virtual human research: face-to-face conversation, emotions and personality, and human figure animation. Assembling a virtual human is still a daunting task, but the building blocks are getting bigger and better every day.

From failures of intelligence analysis to misguided beliefs about vaccinations, biased judgment and decision making contributes to problems in policy, business, medicine, law, education, and private life. Early attempts to reduce decision biases with training met with little success, leading scientists and policy makers to focus on debiasing by using incentives and changes in the presentation and elicitation of decisions. We report the results of two longitudinal experiments that found medium to large effects of one-shot debiasing training interventions. Participants received a single training intervention, played a computer game or watched an instructional video, which addressed biases critical to intelligence analysis (in Experiment 1: bias blind spot, confirmation bias, and fundamental attribution error; in Experiment 2: anchoring, representativeness, and social projection). Both kinds of interventions produced medium to large debiasing effects immediately (games ≥ −31.94% and videos ≥ −18.60%) that persisted at least 2 months later (games ≥ −23.57% and videos ≥ −19.20%). Games that provided personalized feedback and practice produced larger effects than did videos. Debiasing effects were domain general: bias reduction occurred across problems in different contexts, and problem formats that were taught and not taught in the interventions. The results suggest that a single training intervention can improve decision making. We suggest its use alongside improved incentives, information presentation, and nudges to reduce costly errors associated with biased judgments and decisions.