Singapore's Ngee Ann Polytechnic (NP) and London-based Centre for Finance, Technology and Entrepreneurship (CFTE) will jointly launch an industry-led AI in Finance (AIF) online course on June 24, 2018. Through this course, both NP and CFTE hope to support and to nurture talent in Fintech and to boost Fintech development in their respective regions and around the world. The course is accredited by SkillsFuture Singapore and is in the process of obtaining accreditation with The Institute of Banking and Finance Singapore. It aims to update finance professionals and technologists on the AI revolution and create an online community of learners and experts in AI to connect and network for future collaborations. Over 20 finance and technology thought leaders and insiders will come together to share key fundamentals and real-life case studies on how AI is reshaping the finance industry worldwide.

As the uses of Artificial Intelligence (AI) technology expand into every industry, companies like yours might be looking to hire experienced AI and, more specifically, machine learning (ML) talent from a small pool of available candidates stateside. But how can you find, entice, and recruit ML talent when you're competing with major tech giants like Google, Amazon, and Microsoft? The answer lies in getting creative with your recruitment and hiring strategies. The first thing to know when reevaluating recruitment strategies for high-end machine learning (ML) or other AI roles is that you'll need to adapt strategies based on the experience level you're looking for. What works for a Jr. ML Engineer won't work for recruiting Sr. AI Researcher. To access the talent you're looking to hire, you need to go where they'll be found.

The Statistics and Data Science Center (SDSC) at the Massachusetts Institute of Technology is launching a new online micromasters in statistics and data science. Currently under development by MIT faculty, the program will be available through edX in the fall and will feature a curriculum covering foundational knowledge of data science's methods and tools, in-depth coverage of probability and statistics and opportunities to experiment with data analysis techniques and machine learning algorithms. "The demand for data scientists is growing rapidly," said Krishna Rajagopal, dean for digital learning, in a prepared statement. "This new program increases the supply of professionals who are masters of the data science of today, and who have the foundational understanding needed to keep on top of the data science of tomorrow." MIT's micromasters programs are open to anyone who wants to enroll with no application process.

Singapore, South Korea, and Germany topped a recent survey of how countries across the world are dealing with rapidly advancing artificial intelligence (AI) technology primed to automate millions of jobs in the coming years. Commissioned by ABB, The Economist Intelligence Unit created their Automation Readiness Index using data and interviews with industry stakeholders, economists, government officials, and NGOs. The index ranks countries by how prepared they are for what they call "intelligent automation." There have long been questions about the future of automation and the adoption of AI-backed robotics that can do more than just replicate physical human tasks. Businesses are showing little hesitation in exploring AI, but governments continue to struggle in preparing populations for the coming changes to the workforce.

Enrollment in the society, founded in 1946, is open only to individuals who score in the 98th percentile or higher on a pre-approved intelligence test. To join this clever club, hopefuls must demonstrate excellence in verbal-, spatial-, and mathematical-reasoning skills. These four problems are what today's aspirants might see on a typical IQ assessment.

Sony Corp. and Carnegie Mellon University aim to jointly develop a home-use robot for food preparation that utilizes artificial intelligence technology. Sony and the U.S. school have concluded an agreement to collaborate on AI and robotics research. Their initial research and development efforts will focus on optimizing food preparation, cooking and delivery. They want to release a product within five years. "The technology necessary for a robot to handle the complex and varied task of food preparation and delivery could be applied to a broader set of skills and industries," Sony said in a statement.

There are a lot of neat things going on in deep reinforcement learning. One of the coolest things from last year was OpenAI and DeepMind's work on training an agent using feedback from a human rather than a classical reward signal. There's a great blog post about it at Learning from Human Preferences, and the original paper is at Deep Reinforcement Learning from Human Preferences. I've seen a few recommendations that reproducing papers is a good way of levelling up machine learning skills, and I decided this could be an interesting one to try with. It was indeed a super fun project, and I'm happy to have tackled it - but looking back, I realise it wasn't exactly the experience I thought it would be. If you're thinking about reproducing papers too, here are some notes on what surprised me about working with deep RL.

Existing applications include a huge amount of knowledge that is out of reach for deep neural networks. This paper presents a novel approach for integrating calls to existing applications into deep learning architectures. Using this approach, we estimate each application's functionality with an estimator, which is implemented as a deep neural network (DNN). The estimator is then embedded into a base network that we direct into complying with the application's interface during an end-to-end optimization process. At inference time, we replace each estimator with its existing application counterpart and let the base network solve the task by interacting with the existing application. Using this 'Estimate and Replace' method, we were able to train a DNN end-to-end with less data and outperformed a matching DNN that did not interact with the external application.

Many recently trained neural networks employ large numbers of parameters to achieve good performance. One may intuitively use the number of parameters required as a rough gauge of the difficulty of a problem. But how accurate are such notions? How many parameters are really needed? In this paper we attempt to answer this question by training networks not in their native parameter space, but instead in a smaller, randomly oriented subspace. We slowly increase the dimension of this subspace, note at which dimension solutions first appear, and define this to be the intrinsic dimension of the objective landscape. The approach is simple to implement, computationally tractable, and produces several suggestive conclusions. Many problems have smaller intrinsic dimensions than one might suspect, and the intrinsic dimension for a given dataset varies little across a family of models with vastly different sizes. This latter result has the profound implication that once a parameter space is large enough to solve a problem, extra parameters serve directly to increase the dimensionality of the solution manifold. Intrinsic dimension allows some quantitative comparison of problem difficulty across supervised, reinforcement, and other types of learning where we conclude, for example, that solving the inverted pendulum problem is 100 times easier than classifying digits from MNIST, and playing Atari Pong from pixels is about as hard as classifying CIFAR-10. In addition to providing new cartography of the objective landscapes wandered by parameterized models, the method is a simple technique for constructively obtaining an upper bound on the minimum description length of a solution. A byproduct of this construction is a simple approach for compressing networks, in some cases by more than 100 times.

The state of the art in video understanding suffers from two problems: (1) The major part of reasoning is performed locally in the video, therefore, it misses important relationships within actions that span several seconds. (2) While there are local methods with fast per-frame processing, the processing of the whole video is not efficient and hampers fast video retrieval or online classification of long-term activities. In this paper, we introduce a network architecture that takes long-term content into account and enables fast per-video processing at the same time. The architecture is based on merging long-term content already in the network rather than in a post-hoc fusion. Together with a sampling strategy, which exploits that neighboring frames are largely redundant, this yields high-quality action classification and video captioning at up to 230 videos per second, where each video can consist of a few hundred frames. The approach achieves competitive performance across all datasets while being 10x to 80x faster than state-of-the-art methods.