Source coding is the canonical problem of data compression in information theory. In a {\em locally encodable} source coding, each compressed bit depends on only few bits of the input. In this paper, we show that a recently popular model of semisupervised clustering is equivalent to locally encodable source coding. In this model, the task is to perform multiclass labeling of unlabeled elements. At the beginning, we can ask in parallel a set of simple queries to an oracle who provides (possibly erroneous) binary answers to the queries. The queries cannot involve more than two (or a fixed constant number $\Delta$ of) elements. Now the labeling of all the elements (or clustering) must be performed based on the (noisy) query answers. The goal is to recover all the correct labelings while minimizing the number of such queries. The equivalence to locally encodable source codes leads us to find lower bounds on the number of queries required in variety of scenarios. We are also able to show fundamental limitations of pairwise `same cluster' queries - and propose pairwise AND queries, that provably performs better in many situations.

This paper describes multichannel speech enhancement for improving automatic speech recognition (ASR) in noisy environments. Recently, the minimum variance distortionless response (MVDR) beamforming has widely been used because it works well if the steering vector of speech and the spatial covariance matrix (SCM) of noise are given. To estimating such spatial information, conventional studies take a supervised approach that classifies each time-frequency (TF) bin into noise or speech by training a deep neural network (DNN). The performance of ASR, however, is degraded in an unknown noisy environment. To solve this problem, we take an unsupervised approach that decomposes each TF bin into the sum of speech and noise by using multichannel nonnegative matrix factorization (MNMF). This enables us to accurately estimate the SCMs of speech and noise not from observed noisy mixtures but from separated speech and noise components. In this paper we propose online MVDR beamforming by effectively initializing and incrementally updating the parameters of MNMF. Another main contribution is to comprehensively investigate the performances of ASR obtained by various types of spatial filters, i.e., time-invariant and variant versions of MVDR beamformers and those of rank-1 and full-rank multichannel Wiener filters, in combination with MNMF. The experimental results showed that the proposed method outperformed the state-of-the-art DNN-based beamforming method in unknown environments that did not match training data.

We explore deep Reinforcement Learning(RL) algorithms for scalping trading and knew that there is no appropriate trading gym and agent examples. Thus we propose gym and agent like Open AI gym in finance. Not only that, we introduce new RL framework based on our hybrid algorithm which leverages between supervised learning and RL algorithm and uses meaningful observations such order book and settlement data from experience watching scalpers trading. That is very crucial information for traders behavior to be decided. To feed these data into our model, we use spatio-temporal convolution layer, called Conv3D for order book data and temporal CNN, called Conv1D for settlement data. Those are preprocessed by episode filter we developed. Agent consists of four sub agents divided to clarify their own goal to make best decision. Also, we adopted value and policy based algorithm to our framework. With these features, we could make agent mimic scalpers as much as possible. In many fields, RL algorithm has already begun to transcend human capabilities in many domains. This approach could be a starting point to beat human in the financial stock market, too and be a good reference for anyone who wants to design RL algorithm in real world domain. Finally, weexperiment our framework and gave you experiment progress.

Benefitted from its great success on many tasks, deep learning is increasingly used on low-computational-cost devices, e.g. smartphone, embedded devices, etc. To reduce the high computational and memory cost, in this work, we propose a fully learnable group convolution module (FLGC for short) which is quite efficient and can be embedded into any deep neural networks for acceleration. Specifically, our proposed method automatically learns the group structure in the training stage in a fully end-to-end manner, leading to a better structure than the existing pre-defined, two-steps, or iterative strategies. Moreover, our method can be further combined with depthwise separable convolution, resulting in 5 times acceleration than the vanilla Resnet50 on single CPU. An additional advantage is that in our FLGC the number of groups can be set as any value, but not necessarily 2^k as in most existing methods, meaning better tradeoff between accuracy and speed. As evaluated in our experiments, our method achieves better performance than existing learnable group convolution and standard group convolution when using the same number of groups.

Killer robots may sound like the name of a science fiction film, but they could be becoming a reality - and soon. Scientists say artificial intelligence has developed so quickly that we could soon see weapons that can choose a target and kill without being controlled by a human. The United Nations has held five days of talks in Geneva, Switzerland on banning what are known as lethal autonomous weapons. But the United States, Russia, Israel and the United Kingdom are against any restrictions, saying these developments could make war safer. How likely are killer robots?

Digital transformation and modern technologies have impacted every single industry that exists, and the banking industry is no different. In fact, technology has touched upon the banking industry to such a large extent that the union of both has given birth to what is now called'FinTech'. In an interview with DataQuest, Jaya Vaidhyanathan, President, Bahwan CyberTek, gives insights on how new age technologies such as artificial intelligence, machine learning and so on are revolutionizing FinTech, and what more consumers can expect on the same in future. How are new age technologies like artificial intelligence, machine learning, predictive analytics and blockchain transforming the FinTech Space? Technology has impacted the financial sector in such a massive way that it has brought about the union of two industries calling it – Financial Technology (FinTech).

It's not every day that an Air Force captain can give a four-star general an earful. Gen. Stephen Wilson, the vice chief of staff, invites input from his very junior colleague because Kanaan's expertise is artificial intelligence. Wilson says he believes AI's ability to sort mountains of data to find targets like terrorists is a way to change the nature of war. The Air Force needs to lean on Kanaan and other young, tech-savvy airmen, Wilson says, to help transform the way it uses data. "It's pretty unusual," Wilson says of his relationship with Kanaan.

One day in 1992, Lawrence Mattis opened up his mail to find an unsolicited screenplay from two unknown writers. It was a dark, nasty, almost defiantly uncommercial tale of cannibalism and class warfare--the type of story that few execs in Hollywood would want to tell. Yet it was exactly the kind of movie Mattis was looking for. Only a few years earlier, Mattis, in his late twenties, had abandoned a promising legal career to start a talent company, Circle of Confusion, with the aim of discovering new writers to represent. He'd set up shop in New York City, despite being told repeatedly that his best hope for finding talent was to be in Los Angeles. Before that strange script showed up, Mattis was starting to wonder if those naysayers had been right. "I'd only sold a few options that paid about five hundred dollars each," Mattis says. "I was starting to think about going back to law. Then I get this letter from these two kids, saying'Could you please read our script?'" The screenplay, titled Carnivore, was a horror tale set in a soup kitchen, where the bodies of the rich are used to feed the poor. "It was funny, it was visceral, and it made it clear that whoever wrote it really knew movies," Mattis says. Its writers were Lilly and Lana Wachowski, two self-described "schmoes from Chicago" who, in later years, would be referred to by many colleagues and admirers simply as "the Wachowskis." By the time they contacted Mattis, the Wachowskis had been collaborating for years, having spent their childhood creating radio plays, comic books, and their own role-playing game. They'd been raised in a middle-class neighborhood on Chicago's South Side by their mother, a nurse and artist, and their father, a businessman. Growing up, their parents had encouraged them to discover art, especially film.

Samuel Smiths pubs have banned people from using phones, in an attempt to foster "social conversations". Various pubs across the country have been sent a memo making clear that people should not be allowed to use their phones or chat in the bar, and that if they wish to they should be directed outside in the same way as if they were smoking. It follows a similar rule that saw swearing banned in the pubs. The strange rules are said to be enforced by the company's 73-year-old owner, who is rumoured to enter the pubs undercover to check that they are being kept to. We'll tell you what's true.

New pictures could show the inside of the new iPhone – and seem to boost the idea it will have a very complicated camera set up. Apple is said to be readying a whole host of phones for release in September. And one is one of the strangest rumours to surround the iPhone: that it could have at least three lenses in its camera, though it is still not entirely clear what for. New schematics seem to show that new phone, with all of its three lenses. It chimes with previous rumours about that set-up, which have come from leaks from inside Apple's supply chain. We'll tell you what's true.