The achievement gap between students who come from different socio-economic backgrounds is a well-known and persistent problem in education. Disparities in achievements between high and low socio-economic groups are evident in children as young as age 3 years and seem to be a problem the world over. Despite pupils' overall attainment scores rising over the last decade or so, the gap between students from different socio-economic groups remains intractably present and widespread. Family finances play a big part amongst the various reasons for this disparity. Pupils from low socio-economic backgrounds (SEBs) are often only able to attend a few, if any of the extra-curricular activities enjoyed by their more affluent peers. Access to good schools for pupils from SEBs is often reduced as is their access to the educational and occupational aspirations which can impact children's academic achievement.

The world is evolving faster than ever thanks to the proliferation of technology brought about by an increasingly connected global market. And while the changes in education system was slow a few decades ago, the course curriculum was not required to be updated frequently. However, with today's fast pace changes in market needs, new curriculum is required almost every few months. This is primarily due to the fact that while technology, knowledge, and the job market can and do change almost overnight, the education system is rigid, often bureaucratic, and in many cases even resistant to change. This ever expanding gap between academia and the real world has resulted in a whole generation of students who while undoubtedly talented, intelligent, and hardworking simply do not have the skillsets required for employment.

Samsung Research, the advanced R&D hub of Samsung Electronics' SET (end-products) business, has ranked first in two of the world's top global artificial intelligence (AI) machine reading comprehension competitions. Samsung Research recently placed first in the MAchine Reading COmprehension (MS MARCO) Competition held by Microsoft (MS), as well as showing the best performance in TriviaQA* hosted by the University of Washington, proving the excellence of its AI algorithm. With intense competition in developing AI technologies globally, machine reading comprehension competitions such as MS MARCO are booming around the world. MS MARCO and TriviaQA are among the actively researched and used machine reading comprehension competitions along with SQuAD of Stanford University and NarrativeQA of DeepMind. Distinguished universities around the world and global AI firms including Samsung are competing in these challenges.

The problem of vocabulary gap among students in the early years of school, and the resulting impact on school success have received significant attention in the past [3-5, 18, 20, 23]. Early introduction of vocabulary through either direct or indirect instruction helps children learn to read well and forms a strong foundation for literacy, which in turn helps children in accelerated reading to learn. Importantly, while reading new texts, children tend to connect the words they are familiar with to the words exposed in the texts; hence, greater and diverse vocabulary leads to better comprehension of the texts being read. Given the enormity of vocabulary in English language (and most languages in general), new word acquisition is an ongoing process for many years, and 1 sometimes is even lifelong for many people. However, the highest rate of vocabulary development happens in the early years, and teachers in elementary schools focus (often in their own ways, since no universally standardized word lists or procedures exist) a nontrivial amount of time in introducing words to children through both direct and implicit instruction.

This paper studies bandit learning problems with delayed feedback, which included multi-armed bandit (MAB) and bandit convex optimization (BCO). Given only function value information (a.k.a. bandit feedback), algorithms for both MAB and BCO typically rely on (possibly randomized) gradient estimators based on function values, and then feed them into well-studied gradient-based algorithms. Different from existing works however, the setting considered here is more challenging, where the bandit feedback is not only delayed but also the presence of its delay is not revealed to the learner. Existing algorithms for delayed MAB and BCO become intractable in this setting. To tackle such challenging settings, DEXP3 and DBGD have been developed for MAB and BCO, respectively. Leveraging a unified analysis framework, it is established that both DEXP3 and DBGD guarantee an ${\cal O}\big( \sqrt{T+D} \big)$ regret over $T$ time slots with $D$ being the overall delay accumulated over slots. The new regret bounds match those in full information settings.

We propose an efficient protocol for decentralized training of deep neural networks from distributed data sources. The proposed protocol allows to handle different phases of model training equally well and to quickly adapt to concept drifts. This leads to a reduction of communication by an order of magnitude compared to periodically communicating state-of-the-art approaches. Moreover, we derive a communication bound that scales well with the hardness of the serialized learning problem. The reduction in communication comes at almost no cost, as the predictive performance remains virtually unchanged. Indeed, the proposed protocol retains loss bounds of periodically averaging schemes. An extensive empirical evaluation validates major improvement of the trade-off between model performance and communication which could be beneficial for numerous decentralized learning applications, such as autonomous driving, or voice recognition and image classification on mobile phones.

Collectively, machine learning (ML) researchers are engaged in the creation and dissemination of knowledge about data-driven algorithms. In a given paper, researchers might aspire to any subset of the following goals, among others: to theoretically characterize what is learnable, to obtain understanding through empirically rigorous experiments, or to build a working system that has high predictive accuracy. While determining which knowledge warrants inquiry may be subjective, once the topic is fixed, papers are most valuable to the community when they act in service of the reader, creating foundational knowledge and communicating as clearly as possible. Recent progress in machine learning comes despite frequent departures from these ideals. In this paper, we focus on the following four patterns that appear to us to be trending in ML scholarship: (i) failure to distinguish between explanation and speculation; (ii) failure to identify the sources of empirical gains, e.g., emphasizing unnecessary modifications to neural architectures when gains actually stem from hyper-parameter tuning; (iii) mathiness: the use of mathematics that obfuscates or impresses rather than clarifies, e.g., by confusing technical and non-technical concepts; and (iv) misuse of language, e.g., by choosing terms of art with colloquial connotations or by overloading established technical terms. While the causes behind these patterns are uncertain, possibilities include the rapid expansion of the community, the consequent thinness of the reviewer pool, and the often-misaligned incentives between scholarship and short-term measures of success (e.g., bibliometrics, attention, and entrepreneurial opportunity). While each pattern offers a corresponding remedy (don't do it), we also discuss some speculative suggestions for how the community might combat these trends.

We consider a system where agents enter in an online fashion and are evaluated based on their attributes or context vectors. There can be practical situations where this context is partially observed, and the unobserved part comes after some delay. We assume that an agent, once left, cannot re-enter the system. Therefore, the job of the system is to provide an estimated score for the agent based on her instantaneous score and possibly some inference of the instantaneous score over the delayed score. In this paper, we estimate the delayed context via an online convex game between the agent and the system. We argue that the error in the score estimate accumulated over $T$ iterations is small if the regret of the online convex game is small. Further, we leverage side information about the delayed context in the form of a correlation function with the known context. We consider the settings where the delay is fixed or arbitrarily chosen by an adversary. Furthermore, we extend the formulation to the setting where the contexts are drawn from some Banach space. Overall, we show that the average penalty for not knowing the delayed context while making a decision scales with $\mathcal{O}(\frac{1}{\sqrt{T}})$, where this can be improved to $\mathcal{O}(\frac{\log T}{T})$ under special setting.

One of the most fundamental problems in network study is community detection. The stochastic block model (SBM) is one widely used model for network data with different estimation methods developed with their community detection consistency results unveiled. However, the SBM is restricted by the strong assumption that all nodes in the same community are stochastically equivalent, which may not be suitable for practical applications. We introduce pairwise covariates-adjusted stochastic block model (PCABM), a generalization of SBM that incorporates pairwise covariate information. We study the maximum likelihood estimates of the coefficients for the covariates as well as the community assignments. It is shown that both the coefficient estimates of the covariates and the community assignments are consistent under suitable sparsity conditions. Spectral clustering with adjustment (SCWA) is introduced to efficiently solve PCABM. Under certain conditions, we derive the error bound of community estimation under SCWA and show that it is community detection consistent. PCABM compares favorably with the SBM or degree-corrected stochastic block model (DCBM) under a wide range of simulated and real networks when covariate information is accessible.

Though yet to become a standard in schools, artificial intelligence in education has been "a thing" since AI's uptick in the 1980s. In many ways, the two seem made for each other. We use education as a means to develop minds capable of expanding and leveraging the knowledge pool, while AI provides tools for developing a more accurate and detailed picture of how the human mind works. AI's digital, dynamic nature also offers opportunities for student engagement that cannot be found in often out-dated textbooks or in the fixed environment of the typical four-walled classroom. In synergistic fashion, they each have the potential to propel the other forward and accelerate the discovery of new learning frontiers and the creation of innovative technologies.