The world is not moving towards a dystopian future just as yet, but one can't deny the effect innovations in the robotics field will have on human life, especially when it comes to a waning workforce as a result of increased use of automation technologies. Experts have already started sounding the warning bells. As per a report published by advisory analyst firm HfS Research in July, the true impact of the emergence of intelligent automation will be felt on the global industry of 15 million IT services and BPO workers, which will see about 1.4 million job losses--a net decrease of 9%--by 2021. In India, the services industry workforce is expected to shrink by 4.8 lakh by 2021, a decline of 14%--HfS Research estimates that the IT services and BPO industry employs about 3.5 million people in the country. Last year, technology research firm Gartner predicted that one in three jobs will be converted to software, robots and smart machines by 2025.

An umbrella term that refers to information systems inspired by biological systems, AI encompasses multiple technologies including machine learning, deep learning, computer vision, natural language processing (NLP), machine reasoning, and strong AI. These technologies have use cases and applications in almost every industry and promise to significantly change existing business models while simultaneously creating new ones. In sizing and forecasting the total global AI market, Tractica has created a taxonomy of 191 real-world use cases for AI, organized into 27 different industry sectors and corresponding with six major technology categories, plus multiple combinations of technologies. Some use cases – such as image recognition, algorithmic securities trading, and healthcare patient data management – have huge scale potential, while others are niche applications. Likewise, a few key industry sectors including consumer products, business services, advertising, and defense applications will drive significant revenue for AI software implementations in addition to AI-driven hardware and service sales, but during the coming decade the technologies will have an effect on almost every conceivable industry sector.

There's no human instinct more basic than speech, and yet, for many people, talking can be taxing. One in 14 working-age Americans suffer from voice disorders that are often associated with abnormal vocal behaviors -- some of which can cause damage to vocal cord tissue and lead to the formation of nodules or polyps that interfere with normal speech production. Unfortunately, many behaviorally-based voice disorders are not well understood. In particular, patients with muscle tension dysphonia (MTD) often experience deteriorating voice quality and vocal fatigue ("tired voice") in the absence of any clear vocal cord damage or other medical problems, which makes the condition both hard to diagnose and hard to treat. But a team from MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL) and Massachusetts General Hospital (MGH) believes that better understanding of conditions like MTD is possible through machine learning.

This is a list of resources I think would be useful for those who are just starting to explore the amazing Machine Learning domain of Computer Science and want to learn more about Neural Networks and their applications. The general idea behind putting these resources together and publishing this list is that when I just started I saw posts with hundreds of links without description and I simply didn't know which of them are worth spending time on. Focusing on most useful ones and giving short summaries instead is a good idea. I am not a Deep Learning expert and everything I wrote down is just my personal experience with these resources, very subjective opinion. In-depth Convolutional Neural Networks course highly recommended if one wants to learn about image recognition, Computer Vision-related problems and so on. The problemset is amazing; it has probably the best numpy tutorial I have ever seen and makes people implement algorithms they saw in lectures in pure Python numpy, which seems to be a great idea as it helps to get better understanding of how everything actually works.

Earlier this summer, humans submitted their selfies and computer scientists submitted their algorithms to be competitors and judges in the first beauty contest judged by artificial intelligence. The six AI judges were trained to evaluate wrinkles, face symmetry, skin color and several other parameters before choosing men and women winners in various age groups ranging from 18 to 69. The photos, which people submitted through an app, had to show the person free of makeup, facial hair and sunglasses (with a facial hair exception being made for those 65 and older). A total of 60,000 people submitted their selfies, and the winners have been chosen. Alex Zhavoronkov, CSO of Youth Laboratories and CEO of Insilico Medicine, the two companies behind Beauty.AI, said he was surprised by the results.

Knowledge bases in the form of ontologies are receiving increasing attention as they allow to clearly represent both the available knowledge, which includes the knowledge in itself and the constraints imposed to it by the domain or the users. In particular, Datalog± ontologies are attractive because of their property of decidability and the possibility of dealing with the massive amounts of data in real world environments; however, as it is the case with many other ontological languages, their application in collaborative environments often lead to inconsistency related issues. In this paper we introduce the notion of incoherence regarding Datalog± ontologies, in terms of satisfiability of sets of constraints, and show how under specific conditions incoherence leads to inconsistent Datalog± ontologies. The main contribution of this work is a novel approach to restore both consistency and coherence in Datalog± ontologies. The proposed approach is based on kernel contraction and restoration is performed by the application of incision functions that select formulas to delete. Nevertheless, instead of working over minimal incoherent/inconsistent sets encountered in the ontologies, our operators produce incisions over non-minimal structures called clusters. We present a construction for consolidation operators, along with the properties expected to be satisfied by them. Finally, we establish the relation between the construction and the properties by means of a representation theorem. Although this proposal is presented for Datalog± ontologies consolidation, these operators can be applied to other types of ontological languages, such as Description Logics, making them apt to be used in collaborative environments like the Semantic Web.

In this dissertation we propose alternative analysis of distributed stochastic gradient descent (SGD) algorithms that rely on spectral properties of the data covariance. As a consequence we can relate questions pertaining to speedups and convergence rates for distributed SGD to the data distribution instead of the regularity properties of the objective functions. More precisely we show that this rate depends on the spectral norm of the sample covariance matrix. An estimate of this norm can provide practitioners with guidance towards a potential gain in algorithm performance. For example many sparse datasets with low spectral norm prove to be amenable to gains in distributed settings. Towards establishing this data dependence we first study a distributed consensus-based SGD algorithm and show that the rate of convergence involves the spectral norm of the sample covariance matrix when the underlying data is assumed to be independent and identically distributed (homogenous). This dependence allows us to identify network regimes that prove to be beneficial for datasets with low sample covariance spectral norm. Existing consensus based analyses prove to be sub-optimal in the homogenous setting. Our analysis method also allows us to find data-dependent convergence rates as we limit the amount of communication. Spreading a fixed amount of data across more nodes slows convergence; in the asymptotic regime we show that adding more machines can help when minimizing twice-differentiable losses. Since the mini-batch results don't follow from the consensus results we propose a different data dependent analysis thereby providing theoretical validation for why certain datasets are more amenable to mini-batching. We also provide empirical evidence for results in this thesis.

Spreading processes are often modelled as a stochastic dynamics occurring on top of a given network with edge weights corresponding to the transmission probabilities. Knowledge of veracious transmission probabilities is essential for prediction, optimization, and control of diffusion dynamics. Unfortunately, in most cases the transmission rates are unknown and need to be reconstructed from the spreading data. Moreover, in realistic settings it is impossible to monitor the state of each node at every time, and thus the data is highly incomplete. We introduce an efficient dynamic message-passing algorithm, which is able to reconstruct parameters of the spreading model given only partial information on the activation times of nodes in the network. The method is generalizable to a large class of dynamic models, as well to the case of temporal graphs.

Gaussian graphical models are widely used to represent conditional dependence among random variables. In this paper, we propose a novel estimator for data arising from a group of Gaussian graphical models that are themselves dependent. A motivating example is that of modeling gene expression collected on multiple tissues from the same individual: here the multivariate outcome is affected by dependencies acting not only at the level of the specific tissues, but also at the level of the whole body; existing methods that assume independence among graphs are not applicable in this case. To estimate multiple dependent graphs, we decompose the problem into two graphical layers: the systemic layer, which affects all outcomes and thereby induces cross- graph dependence, and the category-specific layer, which represents graph-specific variation. We propose a graphical EM technique that estimates both layers jointly, establish estimation consistency and selection sparsistency of the proposed estimator, and confirm by simulation that the EM method is superior to a simple one-step method. We apply our technique to mouse genomics data and obtain biologically plausible results.

An incremental/online state dynamic learning method is proposed for identification of the nonlinear Gaussian state space models. The method embeds the stochastic variational sparse Gaussian process as the probabilistic state dynamic model inside a particle filter framework. Model updating is done at measurement sample rate using stochastic gradient descent based optimization implemented in the state estimation filtering loop. The performance of the proposed method is compared with state-of-the-art Gaussian process based batch learning methods. Finally, it is shown that the state estimation performance significantly improves due to the online learning of state dynamics.