AI Data Science Summit Hybrid Event @Amdocs Raanana and On-Line Broadcast

IGTCloud

Finally we can meet again and we start with a hybrid (Physical & On-line) event on Aug 3rd, at the Amdocs auditorium in Raanana.

Amdocs

<h2>הנרשמים לכנס גם יקבלו גישה להקלטת הוידאו של הכנס, לצפיה לאחר הכנס</h2><h3>Conference registrants will also have access to the conference video recording, for post-conference viewing</h3>

<h1>Abstracts </h1>09:40 - Osher Yadgar, PhD, Head of AI, AmdocsAI for 5G at Amdocs - From better connectivity to innovation platform Today, as the technological stars align, we are ready to enter into the Fourth Industrial Revolution (4IR). 5G powered by Artificial Intelligence (AI) is a paradigm shift that will provide unprecedented mobile communication capabilities in every aspect: -speed and throughput, mobility along with the scale of connected devices, service agility and energy efficiency, and performance and reliability. Such mobile network capabilities represent a fundamental change in the offering of telecommunications service providers from volume and connectivity to innovation and future services. In this talk, I will present the advantages of 5G and the current state of AI. Then, I will show how they come together to provide a new foundation for future technologies in healthcare, Internet of Things (IoT) , smart cities, autonomous cars, and more. 10:00 - Eldad Kepten, PhD, Head of Data Science at Clalit Research InstituteEvaluating COVID-19 Vaccine Effectiveness in a Real World Setting The Clalit Research Institute is responsible for the utilization of Clalit Health Services’ data in order to improve public health. During the ongoing SARS-CoV-2 pandemic, this includes continuous data-based estimation of vaccine effectiveness across the general population. In this talk, I will discuss the pitfalls in the estimation of vaccine effectiveness during an ongoing vaccination campaign in parallel to changing real world conditions. I will then show how our usage of a rolling cohort design and negative controls can address these challenges. Finally, I will touch on unique problems of accurate statistical estimation of vaccine performance in the current fourth (Delta variant) wave.10:20 - Eyal Gruss, PhD. Poet, New-Media Artist and Machine Learning ResearcherArt must be generated - In 1974, Israeli poet David Avidan prophesied that &quot;computer codes will be the authors of the transcripts of the future, and we shall be the guiding-mind and the approving-eye&quot;. Developments in generative deep-learning methods of the past year, including GPT-3, Maxine, DALL-E and CLIP, may mark the onset of such an era, where creators will communicate ideas rather than media. I will discuss the use case of AI for augmenting and empowering human creativity and expressivity.bio: Eyal Gruss is a poet, new-media artist and machine learning researcher. He holds a PhD in physics, and is an IDF Talpiyot graduate. Eyal creates interactive installations, computer generated art, constrained and algorithmic poetry, audio-visual performances, AI art and online interventions. Eyal teaches computational creativity and creates tools to make state-of-the-art techniques more accessible to artists, educators and the public. https://eyalgruss.com10:40 - Aviv Zeevi, PhD, VP Technological Infrastructure Division, The Israeli Innovation AuthorityIsrael AI Program11:50 - Gilli Shama, PhD, Sagy Tuval, Boris Rabinovitch, Amdocs Data IntelligenceBalancing between latency and accuracy – example from NWDAF - AI projects are to be evaluated in multi aspects, including: model results, product functionality, performance, usability, ROI and governance. These aspects may contradict each other, like improving performance by reducing its accuracy. We present technical methods to improve performance including: vectorization, parallelization, just-in-time complier, and model tuning. We also offer an architecture that enables admin to configure the balance between accuracy and performance and to dilute retraining.12:05 - Sara Philpott, Exec Director of Data Management, OpenetAI/ML Demo for NETWORK DATA ANALYTICS FUNCTION (NWDAF) Network Analytics Function (NWDAF) demo investigates the before and after customer and investment impact of two 3GPP defined use cases for Operators. The Network Load use case demonstrates the lack of predictive intelligence on load performance of components can have direct consequence to user’s enjoyment of the service, as well as financial burden. The second use case, User Equipment Mobility examines the unnecessary investment overhead for both signalling and computation that is required without the NWDAF.12:15 - Vasilis Vagias, Senior AI Architect, Cnvrg.ioOperating System for Machine Learning &amp; AI - Cnvrg.io demo - For enterprise data scientists that need to deliver AI-driven results, cnvrg.io is an operating system for AI that gives data scientists the flexibility, and support to deploy high impact models. cnvrg.io will give a glimpse of how you can rapidly develop and deploy solutions at scale using their AI platform and will highlight the key benefits to the enterprise.12:25 - Prof. Gal Chechik, NVIDIA Director of AI Research, leading NVIDIA research in Israel, and a Professor at the Gonda Brain research institute at Bar-Ilan UniversityReasoning about perception AI aims to build systems that interact with their environment, with people and with other agents in the real world. This vision poses hard algorithmic challenges: from generalizing effectively from little or no samples, through adapting to new domains to communicating in ways that are natural to people. I will discuss our recent research thrusts for facing these challenges. First, an approach to model the high-level structure of visual scene. Second, leveraging compositional structures in attribute space to learn from descriptions without any visual samples. Finally, an approach where agents learn new concepts without labels, by using elimination to reason about their environment. Joint work with colleagues at NVIDIA and Bar-Ilan University. 12:45 - Itay Lieder, PhD, R&amp;D team lead, Intel IT AIActionable insights from the content of websites: New methods implemented at Intel For sales and marketing organizations within large enterprises, identifying and understanding new markets, customers and partners is a key challenge. Intel&#x27;s Sales and Marketing Group (SMG) faces similar challenges while growing in new markets and domains and evolving its existing business. In today&#x27;s complex technological and commercial landscape, there is need for intelligent automation supporting a fine-grained understanding of businesses in order to help SMG sift through millions of companies across many geographies and languages and identify relevant directions. We present a system developed in our company that mines millions of public business web pages, and extracts a faceted customer representation. We focus on two key customer aspects that are essential for finding relevant opportunities: industry segments (ranging from broad verticals such as healthcare, to more specific fields such as &#x27;video analytics&#x27;) and functional roles (e.g., &#x27;manufacturer&#x27; or &#x27;retail&#x27;). To address the challenge of labeled data collection, we enrich our data with external information gleaned from Wikipedia, and develop a semi-supervised multi-label, deep learning model that parses customer website texts and classifies them into their respective facets. Our system scans and indexes companies as part of a large-scale knowledge graph that currently holds tens of millions of connected entities with thousands being fetched, enriched and connected to the graph by the hour in real time, and also supports knowledge and insight discovery. In experiments conducted in our company, we are able to significantly boost the performance of sales personnel in the task of discovering new customers and commercial partnership opportunities.13:00 - Ofer Rosenberg, Lead Software Architect of Cloud AI, QualcommCloud AI 100 - High performance and Power Efficient AI inference acceleration for Cloud &amp; Edge AI inference and training is moving from running on traditional HW (CPU, GPU) to running on dedicated acceleration chips (ASICs). This talk presents Qualcomm’s Cloud AI 100, which offers leadership class performance and power efficiency across many applications, from cloud to edge processing, running a wide variety of network types (classification, segmentation, NLP, transformers). In addition, the talk covers AI 100 comprehensive software solution, and introduces the Edge Box reference design. It also includes new benchmark data from MLCommons latest 1.0 inference submission.13:15 - Ofir Farchy, CTO &amp; Co-Founder, Lynx.MDHealthcare AI case studies with Lynx.MD - The basic properties of real-world data are that it is messy, unstructured, has a steep learning curve, and at the same time holds the promise for a vast array of real-world applications for AI and Machine Learning. Healthcare data represents the extreme case, in both the potential positive impact on people’s lives and the challenges in extracting the value from a very complex healthcare system. In this talk, we will briefly cover select cases of healthcare AI applications, getting a glimpse into the challenges and potential rewards in unlocking the value of healthcare data.14:30 - Yonatan Belinkov, PhD, Assistant Professor at the TechnionInvestigating Gender Bias in Language Models Using Causal Mediation Analysis Many interpretation methods for neural models in natural language processing investigate how information is encoded inside hidden representations. However, these methods can only measure whether the information exists, not whether it is actually used by the model. We propose a methodology grounded in the theory of causal mediation analysis for interpreting which parts of a model are causally implicated in its behavior. The approach enables us to analyze the mechanisms which facilitate the flow of information from input to output through various model components, known as mediators. As a case study, we apply this methodology to analyzing gender bias in pre-trained Transformer language models. We study the role of individual neurons and attention heads in mediating gender bias across three datasets designed to gauge a model’s sensitivity to gender bias. Our mediation analysis reveals that gender bias effects are concentrated in specific components of the model that may exhibit highly specialized behavior.14:45 - Prof Reut Tsarfaty, Associate Professor at Bar-Ilan University and the Head of the ONLP labChallenges and solutions in Hebrew NLP While advances reported for English using Deep Neural Networks are unprecedented, reported advances using such models in Hebrew are few and far between. The problem is twofold. First, Hebrew resources available for training NLP models are not at the same order of magnitude as their English counterparts. Second, there are no accepted tasks and benchmarks to evaluate the progress of Hebrew Neural Models on. In this work we aim to remedy both aspects. First, we present AlephBERT, a large neural pre-trained language model (PLM) for Modern Hebrew, which is trained on larger vocabulary and a larger dataset than any Hebrew PLM before. Second, based on AlephBERT we present a new and novel state-of-the-art Hebrew pipeline containing: Segmentation, Part-of-Speech Tagging, full Morphological Tagging, Named-Entity Recognition and Sentiment Analysis. We show that our AlephBERT-powered models present new state-of-the-art results on all tasks, and reflect on the pros and cons for various architectural decisions in the face of the challenges posed by Hebrew, a Morphologically Rich Languages (MRL). We conclude that AlephBERT provides a solid foundation and a single point of entry for the development of diverse Hebrew NLP applications.15:00 - Noam Ordan, PhD, IAHLT NLP Manager, in collaboration with Dr. Omaima Abboud and Uliana ZapasnayaPredicting Grammatical Errors in Non-Native Writing Using Statistical Models Prepositions, along with articles, are estimated to account for 20-50% of all grammatical errors in English made by non-native speakers. The goal of this research is twofold: first, to develop a model to predict English-as-foreign-language (EFL) learners’ use of prepositions, especially erroneous ones; second, to quantify the common-sensical idea that learners opt for “literal translation”, and in fact to estimate the probability students opt for this option. We rely on techniques developed originally for statistical machine translation, with predictions being highly dependent on the speakers’ native language. Using large amounts of parallel data in Arabic and English, we generate a model that calculates the probability of target-language prepositions based on those of the source language. The model indicates that the distribution of translation equivalents is such that (a) there are many possible translations and (b) a very small number of these equivalents cover most of the probability mass, while the other equivalents are part of a long tail of rare events. We hypothesize and demonstrate that EFL learners tend to translate prepositions to their ‘literal’ (i.e. most probable) target-language equivalents.15:15 - Erez Filosof Timnat, Data Science Team Lead at AppsFlyerModeling categorical variables with Bayesian networks Imagine facing a dataset with multiple categorical variables. There are complex dependencies between these variables, encompassing critical insights about the nature of it. In this talk, I argue that the only model that can actually learn all of these dependencies properly - is a Bayesian network. I’ll explain just how it does that, and how we use it for fraud detection at AppsFlyer.15:30 - Yaniv Avrashi, Lior Turgeman / Data Science Team, Amdocs Managed Services’ R&amp;D unitML for IT OperationsAmdocs Case Study From drawing inspiration from NLP &amp; SNA methods to manage event streams through embeddings of the &quot;strangeness&quot; signal to detect cross-stack anomalies, Amdocs Data Scientists will share a glimpse of the ML methodologies being put into action to modernize IT hybrid operations.15:45 - Micha Breakstone, CEO, NeuraLight (Co-founder Chorus.ai) Leveraging AI to digitize neurological careThis talk presents NeuraLight&#x27;s AI-driven platform which integrates multiple objective and standardized digital markers to accelerate and improve drug development, monitoring, and care for patients with neurological disorders. The technology driving the platform includes proprietary Computer Vision &amp; Deep Learning algorithms which automatically extract a host of digital oculometric markers from facial videos captured with a standard smartphone or webcam. These markers serve as an accurate and reliable proxy for currently used endpoints, and taken together form an accurate snapshot of the neurological status. Our innovative approach leverages both signal processing algorithms (eg blind deconvolutions) and standard machine learning algorithms to extract parameters at a 0.1mm resolution to digitize neurological markers.16:00 - Prof. Amir Zeldes, Associate Professor of Computational Linguistics at Georgetown UniversityA Second Wave of Gold Standard Hebrew Treebanks: Challenges and Applications Most Hebrew language technology is based in one or more ways on the over 30 year old newswire content of the MILA Hebrew Treebank - tokenization, segmentation, tagging and more are all informed by this data. Despite some conversions and updates, and especially the Universal Dependencies Hebrew Treebank (UD Hebrew), the underlying data contains many errors, is no longer in sync with developments in other UD languages, and models trained on it do not generalize well to contemporary data, in particular from the Web. In this talk I point out some issues with current Hebrew treebank data, report on work to update it within <a href="https://www.iahlt.org" target="_blank" data-msys-clicktrack="0" rel="nofollow noopener noreferrer">the Israeli Association of Human Language Technologies</a> (IAHLT), and present our ongoing development of Hebrew data in new domains, as well as some NLP results using state of the art transformer-based models for foundational Hebrew NLP tasks.16:15 - Kfir Bar, PhD, Chief Scientist at Basis TechnologyUsing NLP for Diagnosis of Mental Health DisordersNatural language processing (NLP) tools have shown promise for automatically diagnosing mental-health disorders by analyzing written texts and speech transcriptions. In this session, we will survey some of the latest developments in the field, with specific examples of how computational approaches can detect symptoms. We will also share experimental results and discuss their potential for automatically detecting schizophrenia in patients using only speech patterns16:30 - Einat Orr PhD. CEO &amp; Co-founder of Treeverse, Co-creator of lakeFSReproducing Data Has Never Been Easier - In this talk we share how to version data for frictionless experiments and full reproducibility of your datasets and code. We will use lakeFS, an open source tool that delivers Git operations to object storage based data lakes, to easily implement and automate these processes through the use of branches, merges and commits.Bio - Einat Orr is the CEO and Co-founder of Treeverse, the company behind lakeFS, an open source platform that delivers a git-like experience to object-storage based data lakes. She received her PhD. in Mathematics from Tel Aviv University, in the field of optimization in graph theory. Einat previously led several engineering organizations, most recently as CTO at SimilarWeb.

<h1>Jobs Openings</h1><h2>Amdocs</h2><h3><a href="https://asset.amdocs.com/careers/" target="_blank" data-msys-clicktrack="0" rel="nofollow noopener noreferrer">https://asset.amdocs.com/careers/</a></h3><h3>AI positions</h3><h3><a href="https://asset.amdocs.com/careers/" target="_blank" data-msys-clicktrack="0" rel="nofollow noopener noreferrer">https://jobs.amdocs.com/search/?createNewAlert=false&amp;q=AI&amp;locationsearch=israel</a></h3><h2>Intel</h2><h3>AI &amp; Data Science</h3><h3><a href="https://jobs.intel.com/ShowJob/Id/2906951/Data-Scientist-AI-Expert-for-Advanced-Analytics/ " target="_blank" data-msys-clicktrack="0" rel="nofollow noopener noreferrer">https://jobs.intel.com/ShowJob/Id/2906951/Data-Scientist-AI-Expert-for-Advanced-Analytics/ </a></h3><h3>For Students</h3><h3><a href="https://jobs.intel.com/ShowJob/Id/2919671/AI-Data-Scientist-student-for-Advanced-Analytics/" target="_blank" data-msys-clicktrack="0" rel="nofollow noopener noreferrer">https://jobs.intel.com/ShowJob/Id/2919671/AI-Data-Scientist-student-for-Advanced-Analytics/</a></h3><h3></h3>

<h3>For the physical entrance, in accordance with Corona&#x27;s guidelines, please bring a green vaccination approval, possibly with the Ramzor app. Thanks.</h3><h3>לבאים לכנס הפיזי, בהתאם להנחיות קורונה, נא להביא אתכם לכנס אישור תו ירוק, אפשר עם אפליקצית רמזור, תודה</h3>

Our {communityGuidelinesLink} describe the sort of content we prohibit on Eventbrite. If you suspect an event may be in violation, you can report it to us so we can investigate.

To report other categories of prohibited or illegal content, submit {link}

Ticket sales have ended

AI Data Science Summit Hybrid Event @Amdocs Raanana and On-Line Broadcast

More events from IGTCloud

Discover more events from IGTCloud, from Science & Tech to other experiences you might love.

Ticket sales have ended

AI Data Science Summit Hybrid Event @Amdocs Raanana and On-Line Broadcast

More events from IGTCloud

Discover more events from IGTCloud, from Science & Tech to other experiences you might love.

You might also like...

Browse more events with different dates, prices, and formats to find your next great experience.