Goto

Collaborating Authors

toronto


Realizing the Potential of AI Localism by Stefaan G. Verhulst & Mona Sloane

#artificialintelligence

But even by the usual standards, artificial intelligence has had a turbulent run. Is AI a society-renewing hero or a jobs-destroying villain? As always, the truth is not so categorical. At more than 1,000 pages, Thomas Piketty's doorstop sequel to his previous opus, Capital in the Twenty-First Century, does not disappoint. But whether it will fundamentally change the global debate about inequality is an open question.


Baidu's AI Technology Being Used to Combat Coronavirus

#artificialintelligence

On January 6th, the US Centers for Disease Control and Prevention (CDC) notified the public that a flu-like outbreak was propagating in Wuhan City, in the Hubei Province of China. Subsequently, the World Health Organization (WHO) released a similar report on January 9th. While these responses may seem timely, they were slow when compared to an AI company called BlueDot. BlueDot released a report on December 31st, a full week before the CDC released similar information. Even more impressive, BlueDot predicted the Zika outbreak in Florida six months before the first case in 2016.


Hurtful Words: Quantifying Biases in Clinical Contextual Word Embeddings

arXiv.org Machine Learning

In this work, we examine the extent to which embeddings may encode marginalized populations differently, and how this may lead to a perpetuation of biases and worsened performance on clinical tasks. We pretrain deep embedding models (BERT) on medical notes from the MIMIC-III hospital dataset, and quantify potential disparities using two approaches. First, we identify dangerous latent relationships that are captured by the contextual word embeddings using a fill-in-the-blank method with text from real clinical notes and a log probability bias score quantification. Second, we evaluate performance gaps across different definitions of fairness on over 50 downstream clinical prediction tasks that include detection of acute and chronic conditions. We find that classifiers trained from BERT representations exhibit statistically significant differences in performance, often favoring the majority group with regards to gender, language, ethnicity, and insurance status. Finally, we explore shortcomings of using adversarial debiasing to obfuscate subgroup information in contextual word embeddings, and recommend best practices for such deep embedding models in clinical settings.


How Canada is Gaining an Edge in Artificial Intelligence?

#artificialintelligence

Artificial Intelligence these days has become a new key driver of economic growth. It is a significant field in technology right now. While several countries are racing towards AI supremacy, Canada is attracting the world's tech giants that are pouring mammoth amounts in the region. The country is currently in the midst of the AI boom as companies like Microsoft, Facebook, Google, Huawei, among others are spending huge capital on research hubs in Quebec, Ontario and Alberta. Canada is a world research leader and home to extraordinary AI-driven businesses, and has played a vital role in the advancement of AI.


AI Across the World: Top Cities in AI 2020

#artificialintelligence

Considered by some to be the fourth industrial revolution, the capabilities of AI are ever-growing with new personnel, data and financial power pushing it to greater feats each day, week, month and year. With the wealth of data available to us today, the potential of AI is undeniable. With current debates roaring on which sectors will reap the benefits most, including healthcare, finance, education and more, the only certainty is change. In our top cities in AI blog from 2019, we forecasted (with some help from our industry friends), which cities would emerge as tech hubs, so we thought we'd have another go in 2020! Our list, in no particular order, details some of the cities we think will see some great advancements over the next 11 months.


Eddy Travels closes pre-seed round led by Techstars to scale its AI travel assistant – TechCrunch

#artificialintelligence

Eddy Travels, an AI-powered travel assistant bot which can understand text and voice messages, has closed a pre-seed round of around $500,000 led by Techstars Toronto, Practica Capital and Open Circle Capital VC funds from Lithuania, with angel investors from the U.S., Canada, U.K. Launched in November 2018, Eddy Travels claims to have more than 100,000 users worldwide. Travelers can send voice and text messages to the Eddy Travels bot and get personalized suggestions for the best flights. Because of this ease of use, it now gets 40,000 flight searches per month -- tiny compared to the major travels portals, but not bad for a bot that is available on Facebook Messenger, WhatsApp, Telegram, Rakuten Viber, Line and Slack chat apps. The team is now looking to expand into accommodation, car rentals and other travel services. Eddy Travels search is powered by partnerships with Skyscanner and Emirates Airline.


Disease State Prediction From Single-Cell Data Using Graph Attention Networks

arXiv.org Machine Learning

Single-cell RNA sequencing (scRNA-seq) has revolutionized biological discovery, providing an unbiased picture of cellular heterogeneity in tissues. While scRNA-seq has been used extensively to provide insight into health and disease, it has not been used for disease prediction or diagnostics. Graph Attention Networks have proven to be versatile for a wide range of tasks by learning from both original features and graph structures. Here we present a graph attention model for predicting disease state from single-cell data on a large dataset of Multiple Sclerosis (MS) patients. MS is a disease of the central nervous system that is difficult to diagnose. We train our model on single-cell data obtained from blood and cerebrospinal fluid (CSF) for a cohort of seven MS patients and six healthy adults (HA), resulting in 66,667 individual cells. We achieve $\mathbf{92}$ \% accuracy in predicting MS, outperforming other state-of-the-art methods such as a graph convolutional network, random forest, and multi-layer perceptron. Further, we use the learned graph attention model to get insight into the features (cell types and genes) that are important for this prediction. The graph attention model also allow us to infer a new feature space for the cells that emphasizes the difference between the two conditions. Finally we use the attention weights to learn a new low-dimensional embedding which we visualize with PHATE and UMAP. To the best of our knowledge, this is the first effort to use graph attention, and deep learning in general, to predict disease state from single-cell data. We envision applying this method to single-cell data for other diseases.


Research Scientist - AI and/or Machine Learning Job in Toronto, ON Thomson Reuters

#artificialintelligence

We are looking for candidates with innovative approaches and the expertise to combine and improve existing technologies into Cognitive Computing products that will help our professional customers find the answers they are looking for. Do you want to be part of a creative smart team working on leading edge problems along side other research scientists and engineers? We are looking for a passionate Research Scientist with a consistent record of applied innovation, who is interested in taking Artificial Intelligence technologies to the next level. We are looking for a creative individual who can understand complex business problems, and who can develop scalable algorithms and prototypes to demonstrate the art of possible. Applicants should have significant experience in one or more of the following areas: Natural Language Processing, Machine Learning, Question Answering, Text Mining, Information Retrieval, Distributional Semantics, and/or Discourse Modeling.


Deep Transfer Learning for Physiological Signals

arXiv.org Machine Learning

Deep learning is increasingly common in healthcare, yet transfer learning for physiological signals (e.g., temperature, heart rate, etc.) is under-explored. Here, we present a straightforward, yet performant framework for transferring knowledge about physiological signals. Our framework is called PHASE (PHysiologicAl Signal Embeddings). It i) learns deep embeddings of physiological signals and ii) predicts adverse outcomes based on the embeddings. PHASE is the first instance of deep transfer learning in a cross-hospital, cross-department setting for physiological signals. We show that PHASE's per-signal (one for each signal) LSTM embedding functions confer a number of benefits including improved performance, successful transference between hospitals, and lower computational cost.


Welcoming the Dessa Team to Square

#artificialintelligence

We've acquired Dessa, a Toronto-based company building machine learning applications that address significant real-world challenges for all types of businesses. Their team of world-class engineers will immediately bolster our machine learning and artificial intelligence capabilities at Square. Machine learning is a critical field in technology today, and we've expanded our machine learning work at Square over time through both in-house development and acquisitions like Eloquent Labs. The acquisition of Dessa will help us further boost our machine learning abilities, improve our products, and ultimately pass on the benefits to our customers around the world. For example, machine learning technology can help us enhance products in areas like customer engagement, risk management, and more.