Organizations in virtually every industry worldwide are exploring the benefits of Artificial Intelligence and Machine learning. But with this increased use comes some danger. Cybersecurity expert and McAfee CTO, Steve Grobman will explore the dark side of AI. We tend to think that tech is only used for good or by AI experts, but that’s not always true. The low barrier to entry and technologies focused on making AI fail, “adversarial ML”, create intended and unintended negative consequences in the world of AI that we now live in.

Takeaways:

1. How AI and ML can be used to inflict damage and why organizations and consumers need to be aware of it.

2. How does AI’s low barrier to entry present unintended negative consequences for organizations?

3. How will AI become the snake oil of the 21st century?

• What does AI Governance practice look for?
• Who are the leaders in the space and what are they doing that others aren’t?
• Are we actually seeing a techlash from exceptional early-career talent? Or is it overblown?
• How can you create a culture of cross-functional an integrated ethics and compliance?
• Who is accountable for AI good governance in your organisation? And why?
• Should you start incentivising ethical behaviour?
• Creating AI with the human – and the non-tech functions – within the loop. What are the early successes here?

Seemingly overnight, the world we knew has been up ended by the Covid-19 pandemic. Providing virtual self-service is becoming the primary means for customers to obtain the help and support they need from public and private organizations. With inquiry volumes rocketing, artificially intelligent, conversational chatbots are fast becoming the first-line solution.

But scaling a chatbot up to production readiness can often be challenging. It’s therefore critical to consider any conversational AI implementation as part of a wider digital transformation strategy that delivers the necessary resilience and continuity to drive the business forward - instead of treating it as a point solution that fixes an individual pain point.

Based on first-hand experience and real-life examples, Andy will examine conversational AI in action, discuss how it is revolutionizing business processes, and provide the five key steps that are essential for successful implementation.

To deliver the business value promised by AI and ML technologies IT needs to support the move to an industrialized stance.

But as an IT executive this change can be daunting. How can we test, secure and integrate all the new AI/ML tools our Data Scientists are demanding?

How can we leverage our existing platform investments? What skills do the IT teams need? What part will Kubernetes play with these complex stateful deployments? How can we deliver an omni-channel experience when deployments are spread across data centres, clouds and edge?

These questions cross every industry and organizations large and small. During this session, we will review the gaps that often exist between typical business requirements and the common analytical platforms in use today. We will then take a look at how current trends such as cloud operating models, containerization and Kubernetes could play a role in filling capability gaps when adopted successfully and how best to do that in these challenging times. We’ll end by discussing how to best approach the Industrialization process overall and just what the end-state might look like.

Companies and organizations of all sizes and across all industries are turning to AI to solve real-world problems, deliver innovative products and services, and to gain an advantage in an increasingly competitive marketplace. As organizations increase their use of AI, they face two major challenges – data availability and workload scalability. In this session, we will describe how AI workloads and petabytes worth of training data can be seamlessly scaled together across GPU nodes and regions through the use of NVIDIA DeepOps, Kubeflow, Apache Airflow and the NetApp AI Control Plane. These tools facilitate the rapid provisioning of GPU-accelerated data science workspaces and jobs, the efficient maintenance of dataset and model versions at the data science experiment level, and the automation of data ingestion from a variety of sources across disparate environments. The end result is less idle time for GPUs and data scientists, and an increase in data availability leading to more accurate models.

• How can we build processes whereby our data science and machine learning teams don’t overlook domain expertise when leveraging AutoML to automate different stages of the pipeline?
• How to interweave domain expertise and deep understanding of data context to ensure AutoML delivers for your teams
• Leveraging representation engineering to map structured and unstructured data into a custom data architecture
• Avoiding overreliance on single data type when using AutoML.

• Where are the safe bets? Which vendors in the market are leading the way in terms of successful case studies?
• Where is the next big OPEX reducing use case in your business? What are you doing to identify it?
• First things first: does your data science and engineering team have the necessary tools to optimise any current investments?
• Is operational intelligence about your business driven by AI? And if it isn’t, how can you justify the investment to senior management?

Digital innovation is not possible without the tech skills to drive it. Most leaders today don’t have visibility into the technical abilities of their teams, making it impossible to build skills at scale.

Join longtime AI leader and head of learner experiences at Pluralsight, James Aylward for an in-depth conversation about his experience launching and scaling AI products, upskilling for modern technology roles, and staying adaptable to innovate faster.

Intended Audience:

This talk is relevant for any technology leader who is looking to develop or scale AI skills in their organization.

Takeaways:

• Technologists need modern skills as roles and technologies change over time
• AI/ML and adjacent cloud and infrastructure skills evolve quickly; so should your organization
• If you can become adaptable and keep up with the pace of change, you can build transformative products

By the end of this decade, AI will be practicing law and medicine at levels beyond human ability. It will be deeply and irrevocably involved in the production and distribution of our food supply, keeping order in society and running world commerce. It will be the reason we prevent pandemics and make progress on climate change. Our future and that of AI are inseparable. But the shape of this future and our path toward it are not yet set in stone. The default path of leaving it to the whim of Big Tech, like we did with software, is a dangerous option. AI is imminently more powerful and harder to control than software, which is widely recognized for exacerbating inequality, increasing extremism and sewing societal discord. If there is a way to do better with AI, we have a moral obligation to find it because once AI has its grip on humanity, it will be hard to make it let go. This talk is about what we can do to make the world better with AI.

Inversion problems arise in many petro-physical, seismic or more generally oil and gas applications where a physics based forward equation models the behavior of a system. In many cases, especially when the system’s output can be directly observed, the forward model simply exist for the sake of inversion; a technique for estimating a set of parameters, for which the physics based forward model can replicate the system’s observations.

Inverse problems is solved in a number of different ways, among which is the quasi-Newton’s method. The quasi-Newton’s method admits numerical approximations to allow for practical considerations especially in the computation of the Hessian, which is obtained and used along with the Jacobian and step length to find the minimum of a loss function for which the estimated set of parameters is sufficiently close to

However, in spite of the use of heuristics like the Hessian approximation or the estimated step length from an inexact line search formulation or a relaxation of strong Wolfe conditions during the inexact computation of the step length, the computational burden for real-time waveform inversion for either velocity or pressure reconstruction still poses a problem. Compounding the issue, is the reluctance of organizations to migrate from numerical solutions to data science models, at least not entirely.

To address organizational concerns about the fidelity of a complete data science solution, as well as the need for simplified online computational complexity, a hybrid Physics-Data science approach is introduced. This hybrid approach focuses on improving the time complexity of the forward model. The forward model (typically a partial differential equation) has an asymptotic run time of O(n3) (n correlates with the resolution of the solution). For example, a finite element analysis involving 100,000 grid points would invert a 100,000 X 100,000 matrix for each solution of the forward model in each iteration of the inversion algorithm. Since the quasi Newton’s method admits a lot of approximations for the inverse model, we introduce a data science proxy forward model, that can be trained offline to learn the forward equation, and apply it very quickly during online inversion. Consequently, iterations of the inverse model, compute faster, and converges to a solution in a fraction of the original time.

Driving in San Francisco is hard. Teaching an AV to drive the city's chaotic streets is even harder. In fact, an AV driving the streets of SF experiences a challenging situation 46x more often than in suburban areas. One of the roots of this challenge is predicting what other agents on the road are going to do next.

In particular, the long tail of scenarios the car encounters infrequently but are often the most dangerous: U-turns or K-turns, bicycles swerving into traffic, or pedestrians jaywalking unexpectedly.

This session will cover the inner workings of Cruise's Continuous Learning Machine and explore Cruise's approach to building ML solutions that generalize across the long tail of uncommon scenarios:

• Understanding the AV space: what is the prediction problem and why do we use ML to solve it
• Using self-supervised learning to "auto-label" data at massive scale with high throughput
• Addressing dataset imbalance and model generalization challenges by automatically mining generic error situations
• Tying it all together into a continuous learning machine!

Mystique – The Fight Against The Registration Plate Cloner - Deploying and Scaling an Object Detection / Image Classification Model to Process Half a Million Photos a Day

Mystique detects and blurs rego plate of cars on our portal, protecting seller’s privacy against rego plate cloner who would have been able to falsify the rego plate photo and attach it to another vehicle to conduct criminal activities.

The talk will cover:

• How do we build the rego detection/image classification model
• Clever technique to significantly reduce the false positive
• Deploying the tech at scale to process half a million photos a day

• Introduction
• Need of AI for B2B Media & Telecom companies
• Innovations in Content Viewing Experience
• Q & A

• Streaming video quality of experience measurement
• Optimal encoding, delivery and rendering of graphics and animation
• Virtual humans
• Actor placement and compositing (especially motivated by the pandemic)
• Virtual audiences

Leading researchers and medical professionals have been at the forefront of discovering, identifying, isolating and finding a cure for Covid-19. However, there is more to healthcare beyond the pandemic spanning from personalized drug discovery to genome sequencing to image diagnostics to physician scheduling to asset management – use cases that deliver enhanced patient outcomes using the power of AI.

HPE & Nvidia are market leaders in high-performance computing & AI respectively have been busy delivering joint solutions in the Healthcare segment. We have recently launched a suite of AI enabled healthcare solutions that may of interest to you:

• Optimized infrastructure for Cryo-EM & Genomics
• AI-as-a-Service for Researchers
• Accelerated Image Diagnostics
• Hospital Operations, Scheduling & Patient Safety

Join us in this session to learn more on how you can accelerate the adoption of AI within your organization as well as mitigate the cash flow challenges in deploying these solutions.

• Will personalised medicine be achievable without clear regulatory guidance? How can we get state of the art gene therapies into the market as soon as possible?
• Machine learning for drug target interaction prediction
• Matching suitable patients for clinical trials through deep learning – what partnerships and stakeholders need to be involved here? How to overcome data accessibility challenges associated.

• Clinical decisions need to be made quickly, often with little information and no specific guidelines.
• Using our Electronic Health Record from 13 of our hospitals, we gathered demographic, laboratory, clinical, and treatment data of over 14,000 COVID-19 admitted patients
• Models to predict mortality and assess disease severity, predict need for intubation and survival post-intubation, methods for continuous and remote monitoring, as well as NLP on medical notes, will be presented
• New, COVID-19 specific, challenges and considerations for this effort will be discussed

Artificial intelligence (AI) is the catalyst to design new drugs for the most challenging targets. Using AI, medicinal chemists can now virtually screen new molecules for interactions with proteins of interest before pursuing predicted hits as potential drug candidates.

New AI-platforms can screen at an order of magnitude of tens of billions of molecules -- with speed, scale, and success not seen before. With the need for accelerated drug discovery accentuated by the COVID-19 pandemic, many are hopeful these new AI-platforms can play a role. Atomwise, the startup that invented convolutional neural networks for small molecule drug discovery, is currently partnering with fifteen global research teams to pursue broad-spectrum therapies to not only target this pandemic, but potential future outbreaks.

Join Atomwise CEO Dr. Abe Heifets for a Q&A session with University of Manitoba Faculty of Science Professor Jörg Stetefeld and University of Connecticut Professor Dr. James Cole to discuss their collaboration in exploring broad-spectrum therapies for COVID-19 and other coronaviruses, the use case of AI in drug discovery and how access to technology can catalyze innovation in research.

• Computational Phenomics - the systematic study of human phenotypes and diseases models, are crucial to the success of data-driven clinical development
• The future of data-driven drug discovery relies on combining trial-level and individual-level patient data with predictive analytics to expedite and minimize the time in trials by knowing the performance of historical clinical trials.
• Preparing multiple different existing data sources to feed machine intelligence models to identify various “drivers” of parameters associated clinical trial enrollment lifecycle

Alberta, a province in Canada of 4.5 million people, has among the highest rates of opioid prescribing in the world and opioid prescriptions were a key factor in driving the current opioid crisis. Professional bodies have released clinical practice guidelines for safe opioid prescribing. Prescription monitoring programs like Alberta’s Triplicate Prescription Program (TPP), which are mandated to ensure the appropriate use of opioids and reduce their risk, currently use metrics based on these prescribing guidelines.

Currently, no risk calculators are used in practice that quantify the actual risk of hospitalization or death following a dispensation of prescribed opioids. OKAKI is a public health informatics social enterprise that manages Alberta’s prescription monitoring system. OKAKI and its partners undertook a study to use machine learning (ML) with the province’s large health administrative datasets to develop and test a model to predict the 30-day risk of hospitalization, emergency department visit (ED) or death, at the time of an opioid dispensation. A model that performed better than current approaches could potentially be used to support province-wide monitoring and interventions. The presentation will summarize the study context, methods and results, as well as key challenges in implementing this type of prediction model into public health practice.