Machine Learning is the study of methods for programming computers to learn. Computers are applied to a wide range of tasks, and for most of these it is relatively easy for programmers to design and implement the necessary software. However, there are many tasks for which this is difficult or impossible. These can be divided into four general categories. First, there are problems for which there exist no human experts. For example, in modern automated manufacturing facilities, there is a need to predict machine failures before they occur by analyzing sensor readings. Because the machines are new, there are no human experts who can be interviewed by a programmer to provide the knowledge necessary to build a computer system. A machine learning system can study recorded data and subsequent machine failures and learn prediction rules. Second, there are problems where human experts exist, but where they are unable to explain their expertise. This is the case in many perceptual tasks, such as speech recognition, hand-writing recognition, and natural language understanding. Virtually all humans exhibit expert-level abilities on these tasks, but none of them can describe the detailed steps that they follow as they perform them. Fortunately, humans can provide machines with examples of the inputs and correct outputs for these tasks, so machine learning algorithms can learn to map the inputs to the outputs. Third, there are problems where phenomena are changing rapidly. In finance, for example, people would like to predict the future behavior of the stock market, of consumer purchases, or of exchange rates. These behaviors change frequently, so that even if a programmer could construct a good predictive computer program, it would need to be rewritten frequently. A learning program can relieve the programmer of this burden by constantly modifying and tuning a set of learned prediction rules. Fourth, there are applications that need to be customized for each computer user separately. Consider, for example, a program to filter unwanted electronic mail messages. Different users will need different filters. It is unreasonable to expect each user to program his or her own rules, and it is infeasible to provide every user with a software engineer to keep the rules up-to-date. A machine learning system can learn which mail messages the user rejects and maintain the filtering rules automatically. Machine learning addresses many of the same research questions as the fields of statistics, data mining, and psychology, but with differences of emphasis. Statistics focuses on understanding the phenomena that have generated the data, often with the goal of testing different hypotheses about those phenomena. Data mining seeks to find patterns in the data that are understandable by people. Psychological studies of human learning aspire to understand the mechanisms underlying the various learning behaviors exhibited by people (concept learning, skill acquisition, strategy change, etc.).

Machine learning

Probability Distributions.- Linear Models for Regression.- Linear Models for Classification.- Neural Networks.- Kernel Methods.- Sparse Kernel Machines.- Graphical Models.- Mixture Models and EM.- Approximate Inference.- Sampling Methods.- Continuous Latent Variables.- Sequential Data.- Combining Models.

Pattern Recognition and Machine Learning

Digital processing of speech signals

Results of the ICDAR 2015 Robust Reading Competition are presented. A new Challenge 4 on Incidental Scene Text has been added to the Challenges on Born-Digital Images, Focused Scene Images and Video Text. Challenge 4 is run on a newly acquired dataset of 1,670 images evaluating Text Localisation, Word Recognition and End-to-End pipelines. In addition, the dataset for Challenge 3 on Video Text has been substantially updated with more video sequences and more accurate ground truth data. Finally, tasks assessing End-to-End system performance have been introduced to all Challenges. The competition took place in the first quarter of 2015, and received a total of 44 submissions. Only the tasks newly introduced in 2015 are reported on. The datasets, the ground truth specification and the evaluation protocols are presented together with the results and a brief summary of the participating methods.

/pdf/icdar-2015-competition-on-robust-reading-17zul2skj0.pdf

ICDAR 2015 competition on Robust Reading

This report presents the final results of the ICDAR 2013 Robust Reading Competition. The competition is structured in three Challenges addressing text extraction in different application domains, namely born-digital images, real scene images and real-scene videos. The Challenges are organised around specific tasks covering text localisation, text segmentation and word recognition. The competition took place in the first quarter of 2013, and received a total of 42 submissions over the different tasks offered. This report describes the datasets and ground truth specification, details the performance evaluation protocols used and presents the final results along with a brief summary of the participating methods.

/pdf/icdar-2013-robust-reading-competition-3sb9m2ga5i.pdf

ICDAR 2013 Robust Reading Competition

Nearby scalp channels in multi-channel EEG data exhibit high correlation. A question that naturally arises is whether it is required to record signals from all the electrodes in a group of closely spaced electrodes in a typical measurement setup. One could save on the number of channels that are recorded, if it were possible to reconstruct the omitted channels to the accuracy needed for identifying the relevant information (say, spectral content in the signal), required to carry out a preliminary diagnosis. We address this problem from a compressed sensing perspective and propose a measurement and reconstruction scheme. Working with publicly available EEG database, we put our scheme to experiment and illustrate that if it is only a matter of estimating the frequency content of the signal in various EEG bands, then all the channels need not be recorded. We have achieved an average error below 15% between the original and reconstructed signals with respect to estimation of the spectral content in the delta, theta and alpha bands. We have demonstrated that channels in the 10-10 system of electrode placement can be estimated, with an error less than 10% using recordings on the sparser 10-20 system.

Compressed EEG Acquisition with Limited Channels using Estimated Signal Correlation

In this work, we have investigated various style transfer approaches and (i) examined how the stylized reconstruction changes with the change of loss function and (ii) provided a computationally efficient solution for the same. We have used elegant techniques like depth-wise separable convolution in place of convolution and nearest neighbor interpolation in place of transposed convolution. Further, we have also added multiple interpolations in place of transposed convolution. The results obtained are perceptually similar in quality, while being computationally very efficient. The decrease in the computational complexity of our architecture is validated by the decrease in the testing time by 26.1%, 39.1%, and 57.1%, respectively.

Computationally Efficient Approaches for Image Style Transfer.

We present automatic speech recognition (ASR) systems for Tamil and Kannada based on subword modeling to effectively handle unlimited vocabulary due to the highly agglutinative nature of the languages. We explore byte pair encoding (BPE), and proposed a variant of this algorithm named extended-BPE, and Morfessor tool to segment each word as subwords. We have effectively incorporated maximum likelihood (ML) and Viterbi estimation techniques with weighted finite state transducers (WFST) framework in these algorithms to learn the subword dictionary from a large text corpus. Using the learnt subword dictionary, the words in training data transcriptions are segmented to subwords and we train deep neural network ASR systems which recognize subword sequence for any given test speech utterance. The output subword sequence is then post-processed using deterministic rules to get the final word sequence such that the actual number of words that can be recognized is much larger. For Tamil ASR, We use 152 hours of data for training and 65 hours for testing, whereas for Kannada ASR, we use 275 hours for training and 72 hours for testing. Upon experimenting with different combination of segmentation and estimation techniques, we find that the word error rate (WER) reduces drastically when compared to the baseline word-level ASR, achieving a maximum absolute WER reduction of 6.24% and 6.63% for Tamil and Kannada respectively.

Subword Dictionary Learning and Segmentation Techniques for Automatic Speech Recognition in Tamil and Kannada

Using a known speaker-intrinsic normalization procedure, formant data are scaled by the reciprocal of the geometric mean of the first three formant frequencies. This reduces the influence of the talker but results in a distorted vowel space. The proposed speaker-extrinsic procedure re-scales the normalized values by the mean formant values of vowels. When tested on the formant data of vowels published by Peterson and Barney, the combined approach leads to well separated clusters by reducing the spread due to talkers. The proposed procedure performs better than two top-ranked normalization procedures based on the accuracy of vowel classification as the objective measure.

Intrinsic-cum-extrinsic normalization of formant data of vowels

In our earlier work [1, 2], we employed minimum variance distortionless response (MVDR) and MVDR Bauer respectively, as spectral estimation techniques in place of modified-linear prediction in Discrete cosine transform (DCT) based pitch modification [3]. As a general extension, we introduce psychoacoustic characteristics to [1, 2] resulting in Perceptual-MVDR (PMVDR) and PMVDR-Bauer algorithms utilized here for spectral estimation. Further, we employ Bauer method of spectral factorization in our later algorithm since it results in causal inverse filter. These are used to obtain residual signal from pitch synchronous speech frames. The residual signal is resampled using DCT/IDCT depending on the target pitch scale factor. Finally, forward filters realized from the above factorization are used to get pitch modified speech. The modified speech is evaluated subjectively by 10 listeners and mean opinion scores (MOS) are evaluated for pitch factors from 0.5 to 2. Modified bark spectral distortion (MBSD) measure is also employed to evaluate objective performance. We found that the proposed approach has been rated with higher MOS and has achieved lower MBSD than the time domain pitch synchronous overlap [4], modified-LP method [3] and MVDR based methods [1, 2]. Further, we modified the pitch contours of 20 affirmative sentences to sound like interrogative sentences, using the current as well as our earlier algorithms and compared their performance.

A. G. Ramakrishnan

Papers

Compressed EEG Acquisition with Limited Channels using Estimated Signal Correlation

Computationally Efficient Approaches for Image Style Transfer.

Subword Dictionary Learning and Segmentation Techniques for Automatic Speech Recognition in Tamil and Kannada

Intrinsic-cum-extrinsic normalization of formant data of vowels

Perceptual-MVDR based analysis-synthesis of pitch synchronous frames for pitch modification