AI DevWorld -- Main Stage
Tuesday, October 26, 2021
Wednesday, October 27, 2021
FEATURED TALK: (AI): Responsible AI into Practice - Deliver Trust in Artificial Intelligence Solution
AI has been a key driver in innovation in every industry Organizations have ramped up their effort on leveraging AI to gain a competitive advantage. However, AI solution comes with its own challenges and risk, particularly in regulated industries. There have been numerous instances when AI introduced bias. Organizations must use a balanced approach to accelerating the adoption of AI and prioritize AI governance to ensure trust in the AI system. While AI regulation landscape is still evolving, now is the time for organizations to start taking steps to understand and mitigate AI risks. Responsible AI framework provides guidelines around AI governance for building fair, transparent, ethical, and accountable AI solutions. In this session you will learn about how organizations can follow Responsible AI guidelines and operationalize trust in AI solutions by incorporating AI governance throughout the AI/ML life cycle.
A hands-on deep dive on using Apache Pulsar Apache NiFi with Apache MXNet, OpenVino, TensorFlow Lite, and other Deep Learning Libraries on the actual edge devices including Raspberry Pi with Movidius 2, Google Coral TPU and NVidia Jetson Nano. We run deep learning models on the edge devices and send images, capture real-time GPS and sensor data. With our low coding IoT applications providing easy edge routing, transformation, data acquisition and alerting before we decide what data to stream real-time to our data space. These edge applications classify images and sensor readings real-time at the edge and then send Deep Learning results to Apache NiFi for transformation, parsing, enrichment, querying, filtering and merging data to various data stores. https://www.datainmotion.dev/2019/08/updating-machine-learning-models-at.html
Application of AI within image Processing for defect detection and classification which brings significant reduction in time and resources for the manufacturer - use case in semiconductor mask manufacturing
Through an innovative project, reducing CO2 emissions and all other air pollution induced by the mobility in cities by 30% by deploying a solution for a real-time automatic emission-based road traffic micro-regulation, we managed to use the best of AI technologies. Indeed, AI is the key enabler addressing the complexity of real-time analysis of mobility in crossroads and local air pollution with the trend predictions that leads to recommendation to how to regulate road traffic to decrease air pollution and apply these recommendations directly to traffic lights. Using embedded AI at local camera level was instrumental to allow detecting the different road users (vehicle, public transportation, pedestrian, cyclist…) in real-time, while respecting privacy and GDPR, in order to apply mobility strategies for the optimal mobility management with minimum pollution impact. This last part is combining two AI engines with 5 models. This project, [AI]Roads, is an European awarded project and the outcome is tested in some major cities in EU. Beyond technical challenge, we will share some key advantages of combining AI and embedded AI, which might become the mainstream for some applications, and how we offered a scalable solution to a complex problem: the automatic and best trade-off between air pollution and mobility.
Kubeflow is a popular open source project that delivers a composable software foundation for those who need to build and maintain a scalable ML platform with best-in-class KPIs. This presentation and demonstration will review the streamlined ML workflows and simplified operating patterns in Kubeflow 1.4, which is the Community's 11th release since 2018. In this session, Josh Bottum, who is a Kubeflow Community Product Manager, will lead a review of the latest end-to-end machine learning workflows and discuss how market leaders are using Kubeflow to deliver their ML platforms with native Kubernetes efficiencies and portability.
Starting with ML tutorials seems easy. But how do you scale your ML models from detecting cats and dogs to a full scale business ML model?
AI teams invest a lot of rigor in defining new project guidelines. But the same is not true for killing existing projects. In the absence of clear guidelines, teams let infeasible projects drag on for months. By streamlining the process to fail fast on infeasible projects, teams can significantly increase their overall success with AI initiatives. This talk covers how to fail fast on AI projects. AI projects have a lot more unknowns compared to traditional software projects: availability of right datasets, model training to meet required accuracy threshold, fairness and robustness of recommendations in production, and many more.In order to fail fast, we manage AI initiatives as a conversion funnel analogous to marketing and sales funnels. Projects start at the top of the five-stage funnel and can drop off at any stage, either to be temporarily put on ice or permanently suspended and added to the AI graveyard. Each stage of the AI funnel defines a clear set of unknowns to be validated with a list of time-bound success criteria. In the talk, we cover details of the 5-stage funnel and experiences building a fail-fast culture where the AI graveyard is celebrated!
The Agile Metrics are important to track the health of your projects. They help in tracking the project progress. There are other advanced metrics equally important, like Customer Satisfaction, Employee Satisfaction, and Innovation? Tracking these statistics many times is not easy and straightforward.Did you ever think of applying AI (Artificial Intelligence) to measure these and come up with actionable evidence? The AI-powered with NLP (Natural language Processing) and statistical models not just help in getting a good project insight, it can also help in course corrections, and increase the rate of project success. It can help companies to understand their core strengths, weaknesses, and how to position themselves in the market.Rohit will talk and demonstrate how you can digitally transform your Agile Program Management with AI and NLP. How it enables organizations to take proactive measures that not only make projects successful but also help companies stay competitive and thrive in the market.
AI is a term that has been thrown around in the cybersecurity industry for quite some time. The common components typically referenced when talking about AI are Machine learning and Deep Learning, but what are the differences? When it comes to cybersecurity, AI can be a huge leap forward in combating cyberattacks, but not all solutions are the same. If AI could be the silver bullet, why are today's AI solutions not working? Many of the traditional Machine Learning cybersecurity solutions currently available are causing massive operational challenges as they are not adequately combating the ever-evolving and sophisticated threats. Detection and response-based solutions (EDR) are insufficient because they typically can take 10 minutes or more to identify a threat detected in the environment. It takes sub 3 seconds to infect and start encrypting a system; that is why time is of the essence. You have to prevent the infection and damage it can inflict before it takes root, executes, and spreads. One important item of note is the emerging trend of adversarial machine learning being leveraged by cybercriminals; how can this be combated? Executives and security leaders need to start adopting a preventative approach to cybersecurity utilizing the latest in cutting-edge security solutions, which is only made possible through the use of AI and, more importantly, the use of Deep Learning.The great news is that AI technologies are advancing. Deep learning is proven to be the most effective prevention cybersecurity solution to date, resulting in unmatched prevention rates with proven lowest false positive rates. As organizations evaluate new technologies, a firm understanding of the differences, challenges, and benefits of all AI solutions is a must. Therefore, educational advancements in machine learning and deep learning are well warranted.
From search engine results to social media feeds, the applications powered by AI are ubiquitous in our day to day lives. However, there are many dangers of using AI, from amplifying historical biases to making decisions that we cannot interpret. With the rise of AI-based solutions, the need for us to understand the motivation behind these black-box models is imperative. In this session, we explore real scenarios that show the perils of using AI in the wild and understand why simply optimizing for accuracy or performance is not enough. Learn how these risks can be addressed through the use of various techniques throughout the model development and deployment process.
One of the main issues with ML and DL deployment is finding the right way to train and operationalize the model within the company. Serverless approach for deep learning provides simple, scalable, affordable yet reliable architecture. The challenge of this approach is to keep in mind certain limitations in CPU, GPU and RAM, and organize training and inference of your model. My presentation will show how to utilize services like Amazon SageMaker, AWS Batch, AWS Fargate, AWS Lambda, AWS Step Functions and SageMaker Pipelines to organize deep learning workflows. My talk will be beneficial for machine learning engineers and platform engineers.
Thursday, October 28, 2021
NLP is a key component in many data science systems that must understand or reason about text. This hands-on tutorial uses the open-source Spark NLP library to explore advanced NLP in Python. Spark NLP provides state-of-the-art accuracy, speed, and scalability for language understanding by delivering production-grade implementations of some of the most recent research in applied deep learning. It's the most widely used NLP library in the enterprise today. You'll edit and extend a set of executable Python notebooks by implementing these common NLP tasks: named entity recognition, sentiment analysis, spell checking and correction, document classification, and multilingual and multi domain support. The discussion of each NLP task includes the latest advances in deep learning used to tackle it, including the prebuilt use of BERT embeddings within Spark NLP, using tuned embeddings, and 'post-BERT' research results like XLNet, ALBERT, and roBERTa. Spark NLP builds on the Apache Spark and TensorFlow ecosystems, and as such it's the only open-source NLP library that can natively scale to use any Spark cluster, as well as take advantage of the latest processors from Intel and Nvidia. You'll run the notebooks locally on your laptop, but we'll explain and show a complete case study and benchmarks on how to scale an NLP pipeline for both training and inference.
Natural Language Processing(NLP) is an interesting and challenging field. It becomes even more interesting and challenging when we take into consideration more than one human language. when we perform an NLP on a single language there is a possibility that the interesting insights from another human language might be missed out. The interesting and valuable information may be available in other human languages such as Spanish, Chinese, French, Hindi, and other major languages of the world. Also, the information may be available in various formats such as text, images, audio, and video.
In this talk, I will discuss techniques and methods that will help perform NLP tasks on multi-source and multilingual information. The talk begins with an introduction to natural language processing and its concepts. Then it addresses the challenges with respect to multilingual and multi-source NLP. Next, I will discuss various techniques and tools to extract information from audio, video, images, and other types of files using PyScreenshot, SpeechRecognition, Beautiful Soup, and PIL packages. Also, extracting the information from web pages and source code using pytessaract. Next, I will discuss concepts such as translation and transliteration that help to bring the information into a common language format. Once the language is in a common language format it becomes easy to perform NLP tasks. Next, I will explain with the help of a code walkthrough generating a summary from multi-source and multi-lingual information into a specific language using spacy and stanza packages.
1. Introduction to NLP and concepts (05 Minutes)
2. Challenges in Multi source multilingual NLP (02 Minutes)
3. Tools for extracting information from various file formats (04 Minutes)
4. Extract information from web pages and source code (04 Minutes)
5. Methods to convert information into common language format (05 Minutes)
6. code walkthrough for multi-source and multilingual summary generation (10 Minutes)
We will begin with key stats from Gartner and then ask the panel/co-moderators series of questions to initiate the conversation. During the panel, we will also use online polls to engage the attendees. We will also try to answer attendees' questions as well.The covid-19 pandemic has put a lot of strain on the helpdesk because the majority of the organizations had to start working remotely even if they were not ready for it. We will discuss how conversational AI is assisting helpdesks to navigate these challenges.
Too often, “AI-capable” refers to marketing claims instead of practical value add. For this reason, developers tend to be skeptical about AI-driven development. Slapdash application of AI ends up diminishing developer’s creativity and effectiveness. When implemented in inventive, unique ways, AI dramatically improves the productivity of developers and opens up new opportunities for creativity – especially when applied to cloud app development. Beyond the initial development process, AI has the potential to completely transform the entire application lifecycle by eliminating guesswork and repetitive tasks. AI ensures teams are better equipped to manage application dependencies and ensure that regardless of what changes are made, applications never break and are able to seamlessly adapt to inevitable change. AI-supported development democratizes access to advanced tech, making it possible for any IT team – even the lean, mean ones – to build serious apps. Essentially, AI in the DevOps cycle enables developers to shift-left the quality assurance in a more guided and automated way by assisting them at critical phases in the application building process. Instead of finding problems in production, developers are able to identify them while in the midst of the development lifecycle, so they can remain focused on innovating the best solution rather than the intricacies of hand-coding. Pairing AI with visual, model-driven development allows guidance to be both more powerful and less obtrusive and can compress CI/CD pipelines into days or even hours, instead of weeks. As the Head of AI at OutSystems, António has seen firsthand how quickly developers can change their minds after experiencing the speed and creativity AI enables as a complement to traditional development. In this session, he will provide insight on the three most fundamental design decisions regarding integrating AI into an application platform based on OutSystems experience analyzing models based on tens of millions of application graphs and flows, and explore the implications for improving cloud development productivity by 100x. OutSystems serves enterprise customers like Deloitte, which developed a voice to text tool with deep analysis integrated to capture more accurate notes between advisors and their clients.
In this talk, Aparna Dhinakaran, Founder of Arize AI (Ex-Uber ML), will highlight common model failure modes including model drift, data quality issues, performance degradation, etc. The talk will also surface how ML Observability can address these challenges by monitoring for failures, providing tools to troubleshoot and identify the root cause, as well as playing an important part in the feedback loop to improving models. The talk will highlight best practices and share examples from across the industry.
Session will focus on defining Machine Learning (ML) operational models and how enterprises can leverage it through a framework of governance and model risk management to unlock value. Operationalization is essential to realizing the business value of ML models. We will also overlay the paradigm of DevOps on ML lifecycle management including infusing automated validation of model, removing bias and measurement using KPI's. Example framework and architecture of an ML operational model in action will be showcased, including a starter toolkit.
KEYNOTE (AI): Modzy -- Crossing the AI Valley of Death: Deploying and Monitoring Models in Production at Scale
It’s happened again. You built another AI model that will never see the light of day because it won’t make it past the AI “valley of death” – the crossover of model development to model deployment across your enterprise. The handoff between data science and engineering teams is fraught with friction, outstanding questions around governance and accountability, and who is responsible for different parts of the pipeline and process. Even worse? The patchwork approach when building an AI pipeline leaves many organizations open to risks because of a lack of a holistic approach to security and monitoring.Join us to learn about approaches and solutions for configuring a MLOps pipeline that’s right for your organization. You’ll discover why it’s never too early to plan for operationalization of models, regardless of whether your organization has 1, 10, 100, or 1,000 models in production.The discussion will also reveal the merits of an open container specification that allows you to easily package and deploy models in production from everywhere. Finally, new approaches for monitoring model drift and explainability will be revealed that will help manage expectations with business leaders all through a centralized AI software platform called Modzy®.
Anyone building enterprise level machine learning pipelines understands how challenging managing dependencies can be, and that's exactly why Conda works its magic. However, these dependencies can come with security vulnerabilities that are becoming increasingly exploited with malware as hackers target popular open source libraries. In this session, we're cover the most common next generation of cyber attacks, like the cryto-mining typo-squatting on Matplotlib, as well as what tools and best practices you can put into place to protect your MLOps pipelines from cybersecurity attacks.
You know the AI models deployed in production will need to be monitored and updated. It probably does not surprise you that not everyone does so, and that some large bank with thousands of production models doesn’t quite know where all its AI models are, let alone monitor them. But MLOps goes beyond monitoring models to data engineering to driving business objectives. In this session, we will see how Big Tech Cloud and AI players like Azure and AWS enable MLOps today, and what more we can expect to see.