Ariadna Quattoni

Hidden Conditional Random Fields

We present a discriminative latent variable model for classification problems in structured domains where inputs can be represented by a graph of local observations. A hidden-state conditional random field framework learns a set of latent variables conditioned on local features. Observations need not be independent and may overlap in space and time.

Hidden Conditional Random Fields for Gesture Recognition

We introduce a discriminative hidden-state approach for the recognition of human gestures. Gesture sequences often have a complex underlying structure, and models that can incorporate hidden structures have proven to be advantageous for recognition tasks. Most existing approaches to gesture recognition with hidden states employ a Hidden Markov Model or suitable variant (e.g., a factored or coupled state model) to model gesture streams; a significant limitation of these models is the requirement of conditional independence of observations. In addition, hidden states in a generative model are selected to maximize the likelihood of generating all the examples of a given gesture class, which is not necessarily optimal for discriminating the gesture class against other gestures. Previous discriminative approaches to gesture sequence recognition have shown promising results, but have not incorporated hidden states nor addressed the problem of predicting the label of an entire sequence. In this paper, we derive a discriminative sequence model with a hidden state structure, and demonstrate its utility both in a detection and in a multi-way classification formulation. We evaluate our method on the task of recognizing human arm and head gestures, and compare the performance of our method to both generative hidden state and discriminative fully-observable models.

Latent-Dynamic Discriminative Models for Continuous Gesture Recognition

Many problems in vision involve the prediction of a class label for each frame in an unsegmented sequence. In this paper, we develop a discriminative framework for simultaneous sequence segmentation and labeling which can capture both intrinsic and extrinsic class dynamics. Our approach incorporates hidden state variables which model the sub-structure of a class sequence and learn dynamics between class labels. Each class label has a disjoint set of associated hidden states, which enables efficient training and inference in our model. We evaluated our method on the task of recognizing human gestures from unsegmented video streams and performed experiments on three different datasets of head and eye gestures. Our results demonstrate that our model compares favorably to Support Vector Machines, Hidden Markov Models, and Conditional Random Fields on visual gesture recognition tasks.

Transfer learning for image classification with sparse prototype representations

To learn a new visual category from few examples, prior knowledge from unlabeled data as well as previous related categories may be useful. We develop a new method for transfer learning which exploits available unlabeled data and an arbitrary kernel function; we form a representation based on kernel distances to a large set of unlabeled data points. To transfer knowledge from previous related problems we observe that a category might be learnable using only a small subset of reference prototypes. Related problems may share a significant number of relevant prototypes; we find such a concise representation by performing a joint loss minimization over the training sets of related problems with a shared regularization penalty that minimizes the total number of prototypes involved in the approximation. This optimization problem can be formulated as a linear program that can be solved efficiently. We conduct experiments on a news-topic prediction task where the goal is to predict whether an image belongs to a particular news topic. Our results show that when only few examples are available for training a target topic, leveraging knowledge learnt from other topics can significantly improve performance.

Learning Visual Representations using Images with Captions

Current methods for learning visual categories work well when a large amount of labeled data is available, but can run into severe difficulties when the number of labeled examples is small. When labeled data is scarce it may be beneficial to use unlabeled data to learn an image representation that is low-dimensional, but nevertheless captures the information required to discriminate between image categories. This paper describes a method for learning representations from large quantities of unlabeled images which have associated captions; the goal is to improve learning in future image classification problems. Experiments show that our method significantly outperforms (1) a fully-supervised baseline model, (2) a model that ignores the captions and learns a visual representation by performing PCA on the unlabeled images alone and (3) a model that uses the output of word classifiers trained using captions and unlabeled data. Our current work concentrates on captions as the source of meta-data, but more generally other types of meta-data could be used.

A Joint Model for 2D and 3D Pose Estimation from a Single Image

We introduce a novel approach to automatically recover 3D human pose from a single image. Most previous work follows a pipelined approach: initially, a set of 2D features such as edges, joints or silhouettes are detected in the image, and then these observations are used to infer the 3D pose. Solving these two problems separately may lead to erroneous 3D poses when the feature detector has performed poorly. In this paper, we address this issue by jointly solving both the 2D detection and the 3D inference problems. For this purpose, we propose a Bayesian framework that integrates a generative model based on latent variables and discriminative 2D part detectors based on HOGs, and perform inference using evolutionary algorithms. Real experimentation demonstrates competitive results, and the ability of our methodology to provide accurate 2D and 3D pose estimations even when the 2D detectors are inaccurate.

Spectral learning of weighted automata

In recent years we have seen the development of efficient provably correct algorithms for learning Weighted Finite Automata (WFA). Most of these algorithms avoid the known hardness results by defining parameters beyond the number of states that can be used to quantify the complexity of learning automata under a particular distribution. One such class of methods are the so-called spectral algorithms that measure learning complexity in terms of the smallest singular value of some Hankel matrix. However, despite their simplicity and wide applicability to real problems, their impact in application domains remains marginal to this date. One of the goals of this paper is to remedy this situation by presenting a derivation of the spectral method for learning WFA that—without sacrificing rigor and mathematical elegance—puts emphasis on providing intuitions on the inner workings of the method and does not assume a strong background in formal algebraic methods. In addition, our algorithm overcomes some of the shortcomings of previous work and is able to learn from statistics of substrings. To illustrate the approach we present experiments on a real application of the method to natural language parsing.

A spectral learning algorithm for Finite State Transducers

Finite-State Transducers (FSTs) are a popular tool for modeling paired input-output sequences, and have numerous applications in real-world problems. Most training algorithms for learning FSTs rely on gradient-based or EM optimizations which can be computationally expensive and suffer from local optima issues. Recently, Hsu et al. [13] proposed a spectral method for learning Hidden Markov Models (HMMs) which is based on an Observable Operator Model (OOM) view of HMMs. Following this line of work we present a spectral algorithm to learn FSTs with strong PAC-style guarantees. To the best of our knowledge, ours is the first result of this type for FST learning. At its core, the algorithm is simple, and scalable to large data sets. We present experiments that validate the effectiveness of the algorithm on synthetic and real data.

Load-Sensitive Routing with Directional Antennas

We present a load-sensitive routing (LSR) protocol for static wireless ad hoc networks using directional antennas. This protocol uses the aggregate queue size of all nodes in a path as the load metric for choosing routes. We compare the performance of the load metric with the commonly used hop count metric using directional antennas under different traffic conditions and in different topologies. Data delivery-ratio, end-to-end delay, and throughput are used as the performance metrics to compare the two routing metrics. The load metric performs better than the hop count metric from low to medium traffic rates. At high traffic rates, the performance of both metrics degrade considerably. LSR improves the delivery ratio up to 12%, reduces the end-to-end delay by 250%, and increases the throughput up to 28%

Spectral Learning of Sequence Taggers over Continuous Sequences

In this paper we present a spectral algorithm for learning weighted finite-state sequence taggers (WFSTs) over paired input-output sequences, where the input is continuous and the output discrete. WFSTs are an important tool for modelling paired input-output sequences and have numerous applications in real-world problems. Our approach is based on generalizing the class of weighted finite-state sequence taggers over discrete input-output sequences to a class where transitions are linear combinations of elementary transitions and the weights of the linear combination are determined by dynamic features of the continuous input sequence. The resulting learning algorithm is efficient and accurate.