Schedule: AI case studies sessions

As BuzzFeed’s content production and social networks grow, curation becomes increasingly difficult. To this end, we first built publishing tools that let people work more efficiently. Now, we build artificial intelligence tools that let people work more intelligently. During this talk we plan to share this evolution with the audience. Read more.

Explore how Stripe applies deep-learning to user-behavior for fraud detection. This deep-dive will include data-preparation, modeling methods and performance comparisons. Read more.

Cognitive Solutions, the application of intelligent technology and services to empower the user to draw insights from data using natural human interaction, is a disruptive force in the US Federal market and is changing the way citizens engage with data. Read more.

Twitter is a company with massive amounts of data. Thus, it is no wonder that the company applies machine learning in myriad of ways. In this session, we are going to describe, in depth, one of those use cases: Timeline Ranking. From modeling to infrastructure our goal is to share some of the optimizations that this team have made in order to have models that are both expressive and efficient. Read more.

Overview of data science applications within the asset management industry Specific use cases using ML to derive better investment insights and improve client engagement Read more.

Using AI to combat financial crime is more than strong fraud detection models monitoring transactions. Banks follow significant Anti-Money Laundering (AML) and Know-Your-Customer (KYC) laws and procedures, wrought with growth chained to cost and requiring auditable automation. This session will walk-through a series of case studies that utilize AI-powered RPA that address AML and KYC. Read more.

Drug adverse event narratives contain a wealth of information that is laborious to assess using manual methods. To improve FDA Pharmacovigilance, we apply rule-based text extraction to generate training data for deep learning models. These models improve the identification of adverse events from narrative data, enhance time-to-value, and refine sources of medical terminology. Read more.

We will describe the fundamentals of a next generation AI research project. It is focused on creating future "self-aware" AI systems that have built-in autonomic detection and mitigation facilities to avoid faulty or undesirable behavior in the field: in particular, cognitive bias and inaccurate decisions that are perceived as being unethical. Software-hardware system architectures are discussed. Read more.

Increasingly, companies building modeling platforms to empower their researchers to efficiently scale the development and productionalization of their models. During this talk, we use a case study from a leading algorithmic trading firm to draw general best practices for building these types of platforms in any industry. Read more.

While Deep Learning has shown significant promise towards model performance, it can quickly become untenable particularly when data size is short. RNNs can quickly memorize and over-fit . The presentation explains how a combination of RNNs and Bayesian Network (PGM) can improvise sequence-modeling behavior of RNNs. Read more.

Automated machine learning (AutoML) enables both data scientists and domain experts (with limited machine learning training) to be productive and efficient. AutoML is seen as a fundamental shift in which organizations can approach making machine learning. In this talk, you'll learn how to use auto ML to automate selection of machine learning models and automate tuning of hyper-parameters. Read more.

There is a significant interest in applying deep learning based solutions to problems in medicine and healthcare. This talk will focus on identifying actionable medical problems, and then recasting them as tractable deep learning problems and the techniques to solve them. Read more.

New AI solutions in question answering, chatbots, structured data extraction, text generation, and inference all require deep understanding of the nuances of human language. David Talby shares challenges, risks, and best practices for building NLU-based systems, drawing on examples and case studies from products and services built by Fortune 500 companies and startups over the past seven years. Read more.

Machine learning models are often susceptible to adversarial deception of their input at test time, which is leading to a poorer performance. In this session we will investigate the feasibility of deception in source code attribution techniques in real world environment. This session will present attack scenarios on users identity in open-source projects and discuss possible protection methods. Read more.

Clinical radiology is faced with several clinical issues: 1) improvement in imaging efficiency, 2) reduction of risks, 3) high imaging quality. Subtle Medical provides Deep Learning/AI solution, powered and accelerated by industry solution such as OpenVINO, to address these problems by enabling faster MRI, faster PET and low dose, providing real clinical and financial benefit to hospitals. Read more.

Our firm focuses on the application of AI to investment management. Topics covered in this presentation include the application of AI to the problem of asset selection, dealing with low signal-to-noise ratios in financial time series data, the development of real-time macroeconomic indicators from social media data, and the use of heterogeneous compute architectures, specifically GPUs and FPGAs. Read more.

In this talk we will see how machine learning and deep learning techniques can be applied in the field of quantitative finance. We will look at a few use-cases in detail and see how machine learning techniques can supplement and sometimes even improve upon already existing statistical models. We will also look at novel visualizations to help us better understand and interpret these models. Read more.

A case study that details lights-out automation and how DCL uses AI to transform massive volumes of confidential disparate data into searchable and structured information. Considerations for architecting a solution that processes a continuous flow of 5M+ “pages” of complex work units. Read more.

We propose a framework that unifies Hidden Markov Model and deep learn algorithm (RNN) with modeling components that consider long-term memory and semantics of music (LSTM and Convolution). It takes user original creation as input, modifies the raw scores, and generates musically appropriate melodies. Read more.