How is machine learning changing the world of education? This is a big question. I believe that the application of new technologies from the fields of Machine Learning (ML) and Artificial Intelligence (AI) have the capability of transforming education but that there will be more hype than results in the short run--much like the case of other transformative educational technologies in the past (I am old enough to remember when the filmstrip and, then, the VHS cassette were supposedly going to revolutionize the delivery of instruction in our classrooms). Perhaps the area of promise that has garnered the most attention is "personalized learning." This can be a slippery concept to define (for example, some observers consider Individualized Education Plans as a form of personalized learning--one that requires no particular technology) but I am using it here to mean the use of educational technology to permit students in the same classroom to learn different curricular content, tailored to their own pace and level of mastery (e.g. the Summit Learning platform).

This course is a lead-in to deep learning and neural networks - it covers a popular and fundamental technique used in machine learning, data science and statistics: logistic regression. We cover the theory from the ground up: derivation of the solution, and applications to real-world problems. We show you how one might code their own logistic regression module in Python. This course does not require any external materials. Everything needed (Python, and some Python libraries) can be obtained for free.

The frequency and strength of this user bubble vortex depends on how well your machine learning is set up and the depth of your recommendations, as well as the variety of data points taken into consideration. However, there will always inevitably be this kind of drag in any system where recommendations are based on similarity. That's why it's important to build an element of randomisation into your algorithms. For example, occasionally suggest a piece of content which may not be the optimal match - a bit of a left field choice. This gives users the chance to break free from the bubble and engage with a variety of content.

Courses The major educational initiative of the JHUDSL is to create open-source online courses delivered through a range of platforms including Youtube, Github, Leanpub, and Coursera. Welcome to the "Introduction to Deep Learning" course! In the first week you'll learn about linear models and stochatic optimization methods. Please note that this is an advanced course and we assume basic knowledge of machine learning. I am currently working as a data science researcher and trainee at Jheronimus Academy of Data Science.

After introducing Artificial Intelligence (AI) as an optional skill subject in its schools, the CBSE is now pushing to integrate it with other subjects across all classes. It plans to begin training teachers to this end in the coming months. AI was introduced as an optional skill subject for class IX in the beginning of this year, with the curriculum introducing students to the three domains of AI -- data, computer vision and natural language processing. "…AI is a cognitive science which can be linked to various subjects that concern themselves with cognition and reasoning. Almost every school subject would fall in this domain… It is therefore mandated by the CBSE that all its schools would begin to integrate AI with other disciplines from classes I-XII," read a handbook prepared by the board.

We consider stochastic second order methods for minimizing strongly-convex functions under an interpolation condition satisfied by over-parameterized models. Under this condition, we show that the regularized sub-sampled Newton method (R-SSN) achieves global linear convergence with an adaptive step size and a constant batch size. By growing the batch size for both the sub-sampled gradient and Hessian, we show that R-SSN can converge at a quadratic rate in a local neighbourhood of the solution. We also show that R-SSN attains local linear convergence for the family of self-concordant functions. Furthermore, we analyse stochastic BFGS algorithms in the interpolation setting and prove their global linear convergence. We empirically evaluate stochastic L-BFGS and a "Hessian-free" implementation of R-SSN for binary classification on synthetic, linearly-separable datasets and consider real medium-size datasets under a kernel mapping. Our experimental results show the fast convergence of these methods both in terms of the number of iterations and wall-clock time.

Hierarchical Reinforcement Learning (HRL) is a promising approach to solving long-horizon problems with sparse and delayed rewards. Many existing HRL algorithms either use pre-trained low-level skills that are unadaptable, or require domain-specific information to define low-level rewards. In this paper, we aim to adapt low-level skills to downstream tasks while maintaining the generality of reward design. We propose an HRL framework which sets auxiliary rewards for low-level skill training based on the advantage function of the high-level policy. This auxiliary reward enables efficient, simultaneous learning of the high-level policy and low-level skills without using task-specific knowledge. In addition, we also theoretically prove that optimizing low-level skills with this auxiliary reward will increase the task return for the joint policy. Experimental results show that our algorithm dramatically outperforms other state-of-the-art HRL methods in Mujoco domains. We also find both low-level and high-level policies trained by our algorithm transferable.

Recent studies on catastrophic forgetting during sequential learning typically focus on fixing the accuracy of the predictions for a previously learned task. In this paper we argue that the outputs of neural networks are subject to rapid changes when learning a new data distribution, and networks that appear to "forget" everything still contain useful representation towards previous tasks. Instead of enforcing the output accuracy to stay the same, we propose to reduce the effect of catastrophic forgetting on the representation level, as the output layer can be quickly recovered later with a small number of examples. Towards this goal, we propose an experimental setup that measures the amount of representational forgetting, and develop a novel meta-learning algorithm to overcome this issue. The proposed meta-learner produces weight updates of a sequential learning network, mimicking a multi-task teacher network's representation. We show that our meta-learner can improve its learned representations on new tasks, while maintaining a good representation for old tasks.

We consider the problem of manifold learning. Extending existing approaches of learning from randomly sampled data points, we consider contexts where data may be chosen by a teacher. We analyze learning from teachers who can provide structured data such as points, comparisons (pairs of points), demonstrations (sequences). We prove results showing that the former two do not yield notable decreases in the amount of data required to infer a manifold. Teaching by demonstration can yield remarkable decreases in the amount of data required, if we allow the goal to be teaching up to topology. We further analyze teaching learners in the context of persistence homology. Teaching topology can greatly reduce the number of datapoints required to infer correct geometry, and allows learning from teachers who themselves do not have full knowledge of the true manifold. We conclude with implications for learning in humans and machines.

This paper considers a variant of the classical online learning problem with expert predictions. Our model's differences and challenges are due to lacking any direct feedback on the loss each expert incurs at each time step $t$. We propose an approach that uses peer assessment and identify conditions where it succeeds. Our techniques revolve around a carefully designed peer score function $s()$ that scores experts' predictions based on the peer consensus. We show a sufficient condition, that we call \emph{peer calibration}, under which standard online learning algorithms using loss feedback computed by the carefully crafted $s()$ have bounded regret with respect to the unrevealed ground truth values. We then demonstrate how suitable $s()$ functions can be derived for different assumptions and models.