Balaji Lakshminarayanan

Acoustic classification of multiple simultaneous bird species: A multi-instance multi-label approach

Although field-collected recordings typically contain multiple simultaneously vocalizing birds of different species, acoustic species classification in this setting has received little study so far. This work formulates the problem of classifying the set of species present in an audio recording using the multi-instance multi-label (MIML) framework for machine learning, and proposes a MIML bag generator for audio, i.e., an algorithm which transforms an input audio signal into a bag-of-instances representation suitable for use with MIML classifiers. The proposed representation uses a 2D time-frequency segmentation of the audio signal, which can separate bird sounds that overlap in time. Experiments using audio data containing 13 species collected with unattended omnidirectional microphones in the H. J. Andrews Experimental Forest demonstrate that the proposed methods achieve high accuracy (96.1% true positives/negatives). Automated detection of bird species occurrence using MIML has many potential applications, particularly in long-term monitoring of remote sites, species distribution modeling, and conservation planning.

Non-negative matrix factorization for parameter estimation in hidden Markov models

Hidden Markov models are well-known in analysis of random processes, which exhibit temporal or spatial structure and have been successfully applied to a wide variety of applications such as but not limited to speech recognition, musical scores, handwriting, and bio-informatics. We present a novel algorithm for estimating the parameters of a hidden Markov model through the application of a non-negative matrix factorization to the joint probability distribution of two consecutive observations. We start with the discrete observation model and extend the results to the continuous observation model through a non-parametric approach of kernel density estimation. For both the cases, we present results on a toy example and compare the performance with the Baum-Welch algorithm.

A Syllable-Level Probabilistic Framework for Bird Species Identification

In this paper, we present new probabilistic models for identifying bird species from audio recordings. We introduce the independent syllable model and consider two ways of aggregating frame level features within a syllable. We characterize each syllable as a probability distribution of its frame level features. The independent frame independent syllable (IFIS) model allows us to distinguish syllables whose feature distributions are different from one another. The Markov chain frame independent syllable (MCFIS) model is introduced for scenarios where the temporal structure within the syllable provides significant amount of discriminative information. We derive the Bayes risk minimizing classifier for each model and show that it can be approximated as a nearest neighbour classifier. Our experiments indicate that the IFIS and MCFIS models achieve 88.26% and 90.61% correct classification rates, respectively, while the equivalent SVM implementation achieves 86.15%.

Latent IBP Compound Dirichlet Allocation

We introduce the four-parameter IBP compound Dirichlet process (ICDP), a stochastic process that generates sparse non-negative vectors with potentially an unbounded number of entries. If we repeatedly sample from the ICDP we can generate sparse matrices with an infinite number of columns and power-law characteristics. We apply the four-parameter ICDP to sparse nonparametric topic modelling to account for the very large number of topics present in large text corpora and the power-law distribution of the vocabulary of natural languages. The model, which we call latent IBP compound Dirichlet allocation (LIDA), allows for power-law distributions, both, in the number of topics summarising the documents and in the number of words defining each topic. It can be interpreted as a sparse variant of the hierarchical Pitman-Yor process when applied to topic modelling. We derive an efficient and simple collapsed Gibbs sampler closely related to the collapsed Gibbs sampler of latent Dirichlet allocation (LDA), making the model applicable in a wide range of domains. Our nonparametric Bayesian topic model compares favourably to the widely used hierarchical Dirichlet process and its heavy tailed version, the hierarchical Pitman-Yor process, on benchmark corpora. Experiments demonstrate that accounting for the power-distribution of real data is beneficial and that sparsity provides more interpretable results.

Inference in Supervised latent Dirichlet allocation

Supervised latent Dirichlet allocation (Supervised-LDA) [1] is a probabilistic topic model that can be used for classification. One of the advantages of Supervised-LDA over unsupervised LDA is that it can potentially learn topics that are inline with the class label. The variational Bayes algorithm proposed in [1] for inference in Supervised-LDA suffers from high computational complexity. To address this issue, we develop computationally efficient inference methods for Supervised-LDA. Specifically, we present collapsed variational Bayes and MAP inference for parameter estimation in Supervised-LDA. Additionally, we present computationally efficient inference methods to determine the label of unlabeled data. We provide an empirical evaluation of the classification performance and computational complexity (training as well as classification runtime) of different inference methods for the Supervised-LDA model and a classifier based on probabilistic latent semantic analysis.

Clinically applicable deep learning for diagnosis and referral in retinal disease

The volume and complexity of diagnostic imaging is increasing at a pace faster than the availability of human expertise to interpret it. Artificial intelligence has shown great promise in classifying two-dimensional photographs of some common diseases and typically relies on databases of millions of annotated images. Until now, the challenge of reaching the performance of expert clinicians in a real-world clinical pathway with three-dimensional diagnostic scans has remained unsolved. Here, we apply a novel deep learning architecture to a clinically heterogeneous set of three-dimensional optical coherence tomography scans from patients referred to a major eye hospital. We demonstrate performance in making a referral recommendation that reaches or exceeds that of experts on a range of sight-threatening retinal diseases after training on only 14,884 scans. Moreover, we demonstrate that the tissue segmentations produced by our architecture act as a device-independent representation; referral accuracy is maintained when using tissue segmentations from a different type of device. Our work removes previous barriers to wider clinical use without prohibitive training data requirements across multiple pathologies in a real-world setting.A novel deep learning architecture performs device-independent tissue segmentation of clinical 3D retinal images followed by separate diagnostic classification that meets or exceeds human expert clinical diagnoses of retinal disease.

The Mondrian kernel

We introduce the Mondrian kernel, a fast random feature approximation to the Laplace kernel. It is suitable for both batch and online learning, and admits a fast kernel-width-selection procedure as the random features can be re-used efficiently for all kernel widths. The features are constructed by sampling trees via a Mondrian process [Roy and Teh, 2009], and we highlight the connection to Mondrian forests [Lakshminarayanan et al., 2014], where trees are also sampled via a Mondrian process, but fit independently. This link provides a new insight into the relationship between kernel methods and random forests.