Performance metrics are a key component of machine learning systems, and are ideally constructed to reflect real world tradeoffs. In contrast, much of the literature simply focuses on algorithms for maximizing accuracy. With the increasing integration of machine learning into real systems, it is clear that accuracy is an insufficient measure of performance for many problems of interest. Unfortunately, unlike accuracy, many real world performance metrics are non-decomposable i.e. cannot be computed as a sum of losses for each instance. Thus, known algorithms and associated analysis are not trivially extended, and direct approaches require expensive combinatorial optimization. I will outline recent results characterizing population optimal classifiers for large families of binary and multilabel classification metrics, including such nonlinear metrics as F-measure and Jaccard measure. Perhaps surprisingly, the prediction which maximizes the utility for a range of such metrics takes a simple form. This results in simple and scalable procedures for optimizing complex metrics in practice. I will also outline how the same analysis gives optimal procedures for selecting point estimates from complex posterior distributions for structured objects such as graphs. Joint work with Nagarajan Natarajan, Bowei Yan, Kai Zhong, Pradeep Ravikumar and Inderjit Dhillon.
Organizers: Mijung Park
Creating convincing human facial animation is challenging. Face animation is often hand-crafted by artists separately from body motion. Alternatively, if the face animation is derived from motion capture, it is typically performed while the actor is relatively still. Recombining the isolated face animation with body motion is non-trivial and often results in uncanny results if the body dynamics are not properly reflected on the face (e.g. cheeks wiggling when running). In this talk, I will discuss the challenges of human soft tissue simulation and control. I will then present our method for adding physical effects to facial blendshape animation. Unlike previous methods that try to add physics to face rigs, our method can combine facial animation and rigid body motion consistently while preserving the original animation as closely as possible. Our novel simulation framework uses the original animation as per-frame rest-poses without adding spurious forces. We also propose the concept of blendmaterials to give artists an intuitive means to control the changing material properties due to muscle activation.
Organizers: Timo Bolkart
We transfer a monocular motion stereo 3D reconstruction algorithm from a mobile device (Google Project Tango Tablet) to a rigidly mounted external camera of higher image resolution. A reliable camera synchronization is crucial for the usability of the tablets IMU data and thus a time synchronization method developed. It is based on the joint movement of the cameras. In a second project, we move from outdoor video scenes to aerial images and strive to segment them into polygonal shapes. While most existing approaches address the problem of automated generation of online maps as a pixel-wise segmentation task, we instead frame this problem as constructing polygons representing objects. An approach based on Faster R-CNN, a successful object detection algorithm, is presented.
Organizers: Siyu Tang
We propose a new architecture for the learning of predictive spatio-temporal motion models from data alone. Our approach, dubbed the Dropout Autoencoder LSTM, is capable of synthesizing natural looking motion sequences over long time horizons without catastrophic drift or mo- tion degradation. The model consists of two components, a 3-layer recurrent neural network to model temporal aspects and a novel auto-encoder that is trained to implicitly recover the spatial structure of the human skeleton via randomly removing information about joints during train- ing time. This Dropout Autoencoder (D-AE) is then used to filter each predicted pose of the LSTM, reducing accumulation of error and hence drift over time. Furthermore, we propose new evaluation protocols to assess the quality of synthetic motion sequences even for which no groundtruth data exists. The proposed protocols can be used to assess generated sequences of arbitrary length. Finally, we evaluate our proposed method on two of the largest motion- capture datasets available to date and show that our model outperforms the state-of-the-art on a variety of actions, including cyclic and acyclic motion, and that it can produce natural looking sequences over longer time horizons than previous methods.
Organizers: Gerard Pons-Moll
Estimating 3D shape from monocular 2D images is a challenging and ill-posed problem. Some of these challenges can be alleviated if 3D shape priors are taken into account. In the field of human body shape estimation, research has shown that accurate 3D body estimations can be achieved through optimization, by minimizing error functions on image cues, such as e.g. the silhouette. These methods though, tend to be slow and typically require manual interactions (e.g. for pose estimation). In this talk, we present some recent works that try to overcome such limitations, achieving interactive rates, by learning mappings from 2D image to 3D shape spaces, utilizing data-driven priors, generated from statistically learned parametric shape models. We demonstrate this, either by extracting handcrafted features or directly utilizing CNN-s. Furthermore, we introduce the notion and application of cross-modal or multi-view learning, where abundance of data coming from various views representing the same object at training time, can be leveraged in a semi-supervised setting to boost estimations at test time. Additionally, we show similar applications of the above techniques for the task of 3D garment estimation from a single image.
Organizers: Gerard Pons-Moll
Human observers can classify photographs of real-world scenes after only a very brief exposure to the image (Potter & Levy, 1969; Thorpe, Fize, Marlot, et al., 1996; VanRullen & Thorpe, 2001). Line drawings of natural scenes have been shown to capture essential structural information required for successful scene categorization (Walther et al., 2011). Here, we investigate how the spatial relationships between lines and line segments in the line drawings affect scene classification. In one experiment, we tested the effect of removing either the junctions or the middle segments between junctions. Surprisingly, participants performed better when shown the middle segments (47.5%) than when shown the junctions (42.2%). It appeared as if the images with middle segments tended to maintain the most parallel/locally symmetric portions of the contours. In order to test this hypothesis, in a second experiment, we either removed the most symmetric half of the contour pixels or the least symmetric half of the contour pixels using a novel method of measuring the local symmetry of each contour pixel in the image. Participants were much better at categorizing images containing the most symmetric contour pixels (49.7%) than the least symmetric (38.2%). Thus, results from both experiments demonstrate that local contour symmetry is a crucial organizing principle in complex real-world scenes. Joint work with John Wilder (UofT CS, Psych), Morteza Rezanejad (McGill CS), Kaleem Siddiqi (McGill CS), Allan Jepson (UofT CS), and Dirk Bernhardt-Walther (UofT Psych), to be presented at VSS 2017.
Organizers: Ahmed Osman
Probabilistic deep learning methods have recently made great progress for generative and discriminative modeling. I will give a brief overview of recent developments and then present two contributions. The first is on a generalization of generative adversarial networks (GAN), extending their use considerably. GANs can be shown to approximately minimize the Jensen-Shannon divergence between two distributions, the true sampling distribution and the model distribution. We extend GANs to the class of f-divergences which include popular divergences such as the Kullback-Leibler divergence. This enables applications to variational inference and likelihood-free maximum likelihood, as well as enables GAN models to become basic building blocks in larger models. The second contribution is to consider representation learning using variational autoencoder models. To make learned representations of data useful we need ground them in semantic concepts. We propose a generative model that can decompose an observation into multiple separate latent factors, each of which represents a separate concept. Such disentangled representation is useful for recognition and for precise control in generative modeling. We learn our representations using weak supervision in the form of groups of observations where all samples within a group share the same value in a given latent factor. To make such learning feasible we generalize recent methods for amortized probabilistic inference to the dependent case. Joint work with: Ryota Tomioka (MSR Cambridge), Botond Cseke (MSR Cambridge), Diane Bouchacourt (Oxford)
Organizers: Lars Mescheder
Dynamic events such as family gatherings, concerts or sports events are often photographed by a group of people. The set of still images obtained this way is rich in dynamic content. We consider the question of whether such a set of still images, rather the traditional video sequences, can be used for analyzing the dynamic content of the scene. This talk will describe several instances of this problem, their solutions and directions for future studies. In particular, we will present a method to extend epipolar geometry to predict location of a moving feature in CrowdCam images. The method assumes that the temporal order of the set of images, namely photo-sequencing, is given. We will briefly describe our method to compute photo-sequencing using geometric considerations and rank aggregation. We will also present a method for identifying the moving regions in a scene, which is a basic component in dynamic scene analysis. Finally, we will consider a new vision of developing collaborative CrowdCam, and a first step toward this goal.
Organizers: Jonas Wulff
Deep Learning is one of the most successful machine learning approaches to artificial intelligence. In this talk I discuss the geometry of neural networks as a way to study the success of Deep Learning at a mathematical level and to develop a theoretical basis for making further advances, especially in situations with limited amounts of data and challenging problems in reinforcement learning. I present a few recent results on the representational power of neural networks and then demonstrate how to align this with structures from perception-action problems in order to obtain more efficient learning systems.
Organizers: Jane Walters
Kernel embeddings of distributions and the Maximum Mean Discrepancy (MMD), the resulting distance between distributions, are useful tools for fully nonparametric hypothesis testing and for learning on distributional inputs. I will give an overview of this framework and present some of its recent applications within the context of approximate Bayesian inference. Further, I will discuss a recent modification of MMD which aims to encode invariance to additive symmetric noise and leads to learning on distributions robust to the distributional covariate shift, e.g. where measurement noise on the training data differs from that on the testing data.
Organizers: Philipp Hennig
This talk addresses the task of segmenting moving objects in unconstrained videos. We introduce a novel two-stream neural network with an explicit memory module to achieve this. The two streams of the network encode spatial and temporal features in a video sequence respectively, while the memory module captures the evolution of objects over time. The module to build a “visual memory” in video, i.e., a joint representation of all the video frames, is realized with a convolutional recurrent unit learned from a small number of training video sequences. Given video frames as input, our approach first assigns each pixel an object or background label obtained with an encoder-decoder network that takes as input optical flow and is trained on synthetic data. Next, a “visual memory” specific to the video is acquired automatically without any manually-annotated frames. The visual memory is implemented with convolutional gated recurrent units, which allows to propagate spatial information over time. We evaluate our method extensively on two benchmarks, DAVIS and Freiburg-Berkeley motion segmentation datasets, and show state-of-the-art results. This is joint work with K. Alahari and P. Tokmakov.
Organizers: Osman Ulusoy
Many of the existing Robotics & Automation (R&A) technologies are at a sufficient level of maturity and are widely accepted by the academic (and to a lesser extent by the industrial) community after having undergone the scientific rigor and peer reviews that accompany such works. I believe that most of the past and current research and development efforts in robotics and automation have been squarely aimed at increasing the Standard of Living (SoL) in developed economies where housing, running water, transportation, schools, access to healthcare, to name a few, are taken for granted. Humanitarian R&A, on the other hand, can be taken to mean technologies that can make a fundamental difference in people’s lives by alleviating their suffering in times of need, such as during natural or man-made disasters or in pockets of the population where the most basic needs of humanity are not met, thus improving their Quality of Life (QoL) and not just SoL. My current work focuses on the applied use of robotics and automation technologies for the benefit of under-served and under-developed communities by working closely with them to develop solutions that showcase the effectiveness of R&A solutions in domains that strike a chord with the beneficiaries. This is made possible by bringing together researchers, practitioners from industry, academia, local governments, and various entities such as the IEEE Robotics Automation Society’s Special Interest Group on Humanitarian Technology (RAS-SIGHT), NGOs, and NPOs across the globe. I will share some of my efforts and thoughts on challenges that need to be taken into consideration including sustainability of developed solutions. I will also outline my recent efforts in the technology and public policy domains with emphasis on socio-economic, cultural, privacy, and security issues in developing and developed economies.
Organizers: Ludovic Righetti