In a classroom in rural Maharashtra, students stare at a screen as it reads out their English textbook. The teacher occasionally clicks on a word for its Marathi translation -- the primary language of most students at the school -- or shows them a picture to make them understand a word. This method of learning has started to significantly improve the quality of English education in Maharashtra's government schools. Sanjay Gupta, Chief Executive Officer of EnglishHelper, which created the technology platform, the company developed it after researching the most efficient way to learn a new language. "The challenge is to get exposure to language as a student. This is a multisensory, repetition-based platform which won't confuse the student," he says.

Google's TensorFlow framework is the current leading software for implementing and experimenting with the algorithms that power AI and machine learning. We will embark on this journey by quickly wrapping up some important fundamental concepts, followed by a focus on TensorFlow to complete tasks in computer vision and natural language processing. You will be introduced to some important tips and tricks necessary for enhancing the efficiency of our models. We will highlight how TensorFlow is used in an advanced environment and brush through some of the unique concepts at the cutting edge of practical AI. If you want to develop a solid foundation on using TensorFlow and continue your journey into advancing the state of the art in AI to create your own smart machine learning solutions, this course is for you.

The world of technology is rapidly changing, and one must adapt to it quickly to survive in the race of'survival of the fittest'. The Department of EEE has a team of postgraduate students from the Indian Institute of Technology (IIT) Guwahati, along with their faculty members are developing an Artificial Intelligence-enabled Chatbot named "ALBELA" to teach and support the first-year students of Electrical & Electronics Engineering (EEE). ALBELA is capable of addressing the queries and doubts that each of the approximately 850 students pursuing EEE at the Indian Institute of Technology (IIT) may have and can even scan and convert documents, PDFs or Word and provide results relevant to the query asked. "We have been working on its development since the last 7 months with a team of dedicated 7 research scholars of the department. Earlier we did the trial runs of the Chatbot and started using from this academic session onwards. The response from the students has been overwhelming and we hope that this will become the new normal shortly. Prof. Rohit Sinha, Head EEE Department, and the team IBM have extended their continuous support for this activity."

Deep reinforcement learning(DRL) has been at the center of some of the biggest breakthroughs of artificial intelligence(AI) in the last few years. However, despite all its progress, DRL methods remain incredibly difficult to apply in mainstream solutions given the lack of tooling and libraries. Consequently, DRL remains mostly a research activity that hasn't seen a lot of adoption into real world machine learning solutions. Addressing that problem requires better tools and frameworks. Among the current generation of artificial intelligence(AI) leaders, DeepMind stands alone as the company that has done the most to advance DRL research and development. Recently, the Alphabet subsidiary has been releasing a series of new open source technologies that can help to streamline the adoption of DRL methods.

Automatic analysis of teacher and student interactions could be very important to improve the quality of teaching and student engagement. However, despite some recent progress in utilizing multimodal data for teaching and learning analytics, a thorough analysis of a rich multimodal dataset coming for a complex real learning environment has yet to be done. To bridge this gap, we present a large-scale MUlti-modal Teaching and Learning Analytics (MUTLA) dataset. This dataset includes time-synchronized multimodal data records of students (learning logs, videos, EEG brainwaves) as they work in various subjects from Squirrel AI Learning System (SAIL) to solve problems of varying difficulty levels. The dataset resources include user records from the learner records store of SAIL, brainwave data collected by EEG headset devices, and video data captured by web cameras while students worked in the SAIL products. Our hope is that by analyzing real-world student learning activities, facial expressions, and brainwave patterns, researchers can better predict engagement, which can then be used to improve adaptive learning selection and student learning outcomes. An additional goal is to provide a dataset gathered from the real-world educational activities versus those from controlled lab environments to benefit educational learning community.

A Note on Optimal Sampling Strategy for Structural V ariant Detection Using Optical Mapping Weiwei Li Department of Statistics and Operations Research University of North Carolina at Chapel Hill weiweili@live.unc.edu Abstract Structural variants compose the majority of human genetic variation, but are difficult to assess using current genomic sequencing technologies. Optical mapping technologies, which measure the size of chromosomal fragments between labeled markers, offer an alternative approach. As these technologies mature towards becoming clinical tools, there is a need to develop an approach for determining the optimal strategy for sampling biological material in order to detect a variant at some threshold. Here we develop an optimization approach using a simple, yet realistic, model of the genomic mapping process using a hyper-geometric distribution and probabilistic concentration inequalities. Our approach is both computationally and analytically tractable and includes a novel approach to getting tail bounds of hyper-geometric distribution. We show that if a genomic mapping technology can sample most of the chromosomal fragments within a sample, comparatively little biological material is needed to detect a variant at high confidence. 1 Introduction Structural variants (SV), insertions, deletions, translocations, copy number variants, are by far the most common types of human genetic variation (Chaisson et al., 2015). They have been linked to large number of heritable disorders (Hurles et al., 2008). Technology to assay the presence or absence of these variants has steadily improved in ease and resolution (Huddleston and Eichler, 2016; Audano et al., 2019).

The regression discontinuity (RD) design is one of the most popular quasi-experimental methods for applied causal inference. In practice, the method is quite sensitive to the assumption that individuals cannot control their value of a "running variable" that determines treatment status precisely. If individuals are able to precisely manipulate their scores, then point identification is lost. We propose a procedure for obtaining partial identification bounds in the case of a discrete running variable where manipulation is present. Our method relies on two stages: first, we derive the distribution of non-manipulators under several assumptions about the data. Second, we obtain bounds on the causal effect via a sequential convex programming approach. We also propose methods for tightening the partial identification bounds using an auxiliary covariate, and derive confidence intervals via the bootstrap. We demonstrate the utility of our method on a simulated dataset.

Because of high dimensionality, correlation among covariates, and noise contained in data, dimension reduction (DR) techniques are often employed to the application of machine learning algorithms. Principal Component Analysis (PCA), Linear Discriminant Analysis (LDA), and their kernel variants (KPCA, KLDA) are among the most popular DR methods. Recently, Supervised Kernel Principal Component Analysis (SKPCA) has been shown as another successful alternative. In this paper, brief reviews of these popular techniques are presented first. We then conduct a comparative performance study based on three simulated datasets, after which the performance of the techniques are evaluated through application to a pattern recognition problem in face image analysis. The gender classification problem is considered on MORPH-II and FG-NET, two popular longitudinal face aging databases. Several feature extraction methods are used, including biologically-inspired features (BIF), local binary patterns (LBP), histogram of oriented gradients (HOG), and the Active Appearance Model (AAM). After applications of DR methods, a linear support vector machine (SVM) is deployed with gender classification accuracy rates exceeding 95% on MORPH-II, competitive with benchmark results. A parallel computational approach is also proposed, attaining faster processing speeds and similar recognition rates on MORPH-II. Our computational approach can be applied to practical gender classification systems and generalized to other face analysis tasks, such as race classification and age prediction.

Federated Learning (FL) is currently the most widely adopted framework for collaborative training of (deep) machine learning models under privacy constraints. Albeit it's popularity, it has been observed that Federated Learning yields suboptimal results if the local clients' data distributions diverge. To address this issue, we present Clustered Federated Learning (CFL), a novel Federated Multi-Task Learning (FMTL) framework, which exploits geometric properties of the FL loss surface, to group the client population into clusters with jointly trainable data distributions. In contrast to existing FMTL approaches, CFL does not require any modifications to the FL communication protocol to be made, is applicable to general non-convex objectives (in particular deep neural networks) and comes with strong mathematical guarantees on the clustering quality. CFL is flexible enough to handle client populations that vary over time and can be implemented in a privacy preserving way. As clustering is only performed after Federated Learning has converged to a stationary point, CFL can be viewed as a post-processing method that will always achieve greater or equal performance than conventional FL by allowing clients to arrive at more specialized models. We verify our theoretical analysis in experiments with deep convolutional and recurrent neural networks on commonly used Federated Learning datasets.

Deep neural networks (DNNs) have been shown to over-fit a dataset when being trained with noisy labels for a long enough time. To overcome this problem, we present a simple and effective method self-ensemble label filtering (SELF) to progressively filter out the wrong labels during training. Our method improves the task performance by gradually allowing supervision only from the potentially non-noisy (clean) labels and stops learning on the filtered noisy labels. For the filtering, we form running averages of predictions over the entire training dataset using the network output at different training epochs. We show that these ensemble estimates yield more accurate identification of inconsistent predictions throughout training than the single estimates of the network at the most recent training epoch. While filtered samples are removed entirely from the supervised training loss, we dynamically leverage them via semi-supervised learning in the unsupervised loss. We demonstrate the positive effect of such an approach on various image classification tasks under both symmetric and asymmetric label noise and at different noise ratios. It substantially outperforms all previous works on noise-aware learning across different datasets and can be applied to a broad set of network architectures. The acquisition of large quantities of a high-quality human annotation is a frequent bottleneck in applying DNNs. There are two cheap but imperfect alternatives to collect annotation at large scale: crowdsourcing from non-experts and web annotations, particularly for image data where the tags and online query keywords are treated as valid labels. Both these alternatives typically introduce noisy (wrong) labels. While Rolnick et al. (2017) empirically demonstrated that DNNs can be surprisingly robust to label noise under certain conditions, Zhang et al. (2017) has shown that DNNs have the capacity to memorize the data and will do so eventually when being confronted with too many noisy labels. Consequently, training DNNs with traditional learning procedures on noisy data strongly deteriorates their ability to generalize - a severe problem.