[email protected] conferences: Graph Data Science ran on December 1, 2020.
[email protected] conferences are one-day technology conferences from Manning Publications. Manning authors and other industry experts deliver live coding sessions, deep dives, and tutorials. Watch for free on Twitch.
free one day conference
All [email protected] conferences are free to attend. Just register for your lobby pass.
talks from experts
Featuring four expert speakers, plus ten minute lightning talks.
Missed [email protected] conferences: Graph Data Science? Rewatch now on Manning’s YouTube channel.
conference speakers
Paco Nathan
Paco Nathan is known as a “player/coach”, with core expertise in data science, natural language, cloud computing. He has spent 40 years in tech, with experience ranging from Bell Labs to early-stage start-ups. He has been advisor for Amplify Partners, IBM Data Science Community, Recognai, KUNGFU.AI, and Primer, and is the lead committer for PyTextRank and kglab. Formerly he has been the Director of Community Evangelism at Databricks and Apache Spark. In 2015, he was cited as one of the Top 30 People in Big Data and Analytics by Innovation Enterprise.
Python has several excellent libraries for working with graphs which provide: semantic technologies, graph queries, interactive visualizations, graph algorithms, probabilistic graph inference, as well as embedding and other integrations with deep learning. However, almost none of these have integration paths other than writing lots of custom code, and most do not share common file formats. Moreover, few of these libraries integrate effectively with popular data science tools (e.g., pandas, scikit-learn, PyTorch, spaCy, etc.) or with popular infrastructure for scale-out (Apache Spark, Ray, RAPIDS, Apache Parquet, fsspec, etc.) on cloud computing. This talk shows `kglab` https://github.com/DerwenAI/kglab – an open source project that integrates RDFlib, OWL-RL, pySHACL, NetworkX, iGraph, pslpython, node2vec, PyVis, and more, which speaks fluent PyData and leverages Parquet, Ray, etc., for efficient data engineering with graph data.
Géza Kulcsár
Géza Kulcsár is a senior researcher at IncQuery Labs, a Budapest-based R&D. He holds a PhD in computer science from the Technical University of Darmstadt, Germany. His research interests range from the semantics of graph structures and their transformations to industrial practices in the conceptual modeling of large-scale systems.
There are two fields of growing importance which are very rarely examined together: graph-based modeling and natural language processing (NLP). Graph-based domain-specific knowledge bases such as large-scale system models provide a structured description of engineering data, whose inherent semantics often remains unexplored. In contrast, NLP techniques focus on semantic notions, but do not directly consider the structure of information sources. In this paper, we investigate natural-language querying, i.e., model comprehension based on semantic search for large-scale system models: how do state-of-the-art NLP models and algorithms perform in the presence of a seemingly different semantics? And – is this semantics really that different?
Dhanya Jothimani
Dhanya Jothimani holds a PhD in Financial Analytics from Indian Institute of Technology Delhi, India. She is interested in exploring how data drives the organization and structure of complex systems. Currently, she is working for a leading financial institution in Canada as a Senior Data Scientist.
There has been an interest in investment and stock prices and it has increased over the years. In this talk, we will discuss how the stocks can be related using a graph based approach. Also, I would discuss how these interrelationships between the stocks is (potentially) useful in real life applications.
Luis Guillermo Natera Orozco
Luis Natera is a Ph.D. candidate in Network Science at the Department of Network and Data Science at Central European University in Budapest, Hungary. His research revolves around urban mobility in multiplex urban networks, bridging urban planning and complex systems. Currently, he is working on the application of network science methods for sustainable urban mobility. Luis has experience working in academic, government and private sectors
Urban transportation networks, formed by layers like sidewalks, bicycle paths, or streets, provide the backbone for movement and socio-economic life in cities. To make urban transport sustainable, cities are increasingly investing in their bicycle networks, however, given the patchy nature of the bicycle layer, it is yet unclear how to extend it comprehensively and effectively given a limited budget. In the first part of the talk we will discuss data-driven, algorithmic network growth strategies and how they can be applied to cities worldwide, showing that small but focused investments significantly increase the connectedness and directness of urban bicycle networks. Subsequently, we will show how the pedestrian infrastructure network provides information to compute the liveability of a city, analyzing a European city in terms of walkability. We show novel ways to measure life quality in a more granular scale, using open source datasets, and computational approaches that can be generalized and applyied to any city.
Minko Mihaly & Ágoston Nagy
Mihály Minkó is a data visualization expert and trainer, focusing on dashboard design and network visualization. He often organizes meetups and workshops on different topics that serve as a meeting point for the Hungarian data visualization community. He is the member of the Immersive Media Lab research group at MOME, where one of his focus is the reconstruction of lost phenomena the natural world based on available data. He leads the data visualization training at the same university.
Agoston Nagy is a member of the Binaura artist collective, that has been making algorithmic art and responsive environments using free and open source tools since the early 2000s.
Agoston Nagy is a member of the Binaura artist collective, that has been making algorithmic art and responsive environments using free and open source tools since the early 2000s.
Data visualization has always been considered primarily with the 2D space. Points, lines and areas were its main forms, most of the cases accompanied with color, size or texture. This is the case with network visualizations as well, but their visual vocabulary is even more limited. Lines as edges, circles or other shapes as nodes. Beautiful in it’s simplicity, and it still can convey sometimes more information than any complex chart.
We had the privilege to contribute with two installations to the exhibition of Albert-László Barabási, the well known network scientist. One of them is a series of network visualizations of hungarian artists, who are the first one hundred most “prestigious” – whatever that means in the context of art. The visualizations are drawn by a robot, every day, repetitively, on black paper with white ink, simple, yet compelling. Every day the result is put onto the wall next to the already drawn ones in a special space-filling pattern. The other one resembles a quite different view at the opposite corner of the exhibition area: this is an augmented reality network that overlays on a 3D printed physical statue, that is a food network. Funny, isn’t it? When you print out a construction with a 3D printer, you lose the information richness of the computer screen, the possibilities of interactive surfaces. One way to add that back is to add an augmented layer of information to the physical object, that can show even more information than a screen.
Extending network visualization towards simplicity and complexity at the same time was an absolutely rewarding experience. We are looking forward to creating similarly immersive or meditative experiences to show that there are ways unthought. Yet.
We had the privilege to contribute with two installations to the exhibition of Albert-László Barabási, the well known network scientist. One of them is a series of network visualizations of hungarian artists, who are the first one hundred most “prestigious” – whatever that means in the context of art. The visualizations are drawn by a robot, every day, repetitively, on black paper with white ink, simple, yet compelling. Every day the result is put onto the wall next to the already drawn ones in a special space-filling pattern. The other one resembles a quite different view at the opposite corner of the exhibition area: this is an augmented reality network that overlays on a 3D printed physical statue, that is a food network. Funny, isn’t it? When you print out a construction with a 3D printer, you lose the information richness of the computer screen, the possibilities of interactive surfaces. One way to add that back is to add an augmented layer of information to the physical object, that can show even more information than a screen.
Extending network visualization towards simplicity and complexity at the same time was an absolutely rewarding experience. We are looking forward to creating similarly immersive or meditative experiences to show that there are ways unthought. Yet.
Orsolya Putz
Orsolya Putzholds a PhD in Cognitive Linguistics and devotes herself to cognitive sciences. Currently she is an assistant lecturer at Eötvös Loránd University, Budapestandtheco-founder of Crow Intelligence,a boutique consultancy specialized in NLP and AI.Her main research areas are cognitive metaphor theory, text analytics, and cognitive background of biases in human and machine models. Shealso worked as a linguistic expert on various text analytics projects.
The talk aims to present how to build similarity graphs of different types of datasets (e.g. texts and images). It will be shown that the same method can be applied to multiple modalities. The resulting graphs will be simplified with backbone filtering. Finally, it will be demonstrated that the simplified graphs can be used for various content analysis tasks (e.g. extractive text summarization and identification of prototypical images).
András Vicsek
CEO at Maven7
Lynx Analytics is a pioneer in driving business value from graph based insights. Our international clientele range from large telcos through banks to transport agencies and governments. Our projects vary from graph based advanced demography estimation, customer satisfaction monitoring through viral campaign design to fiber network and ATM location optimization. The common denominator for all these projects is graph: we always beat state of the art by somehow using the extra information available in some form of relationship data.
One of the important sources of Lynx’s success is LynxKite. LynxKite is a graph data science platform. It is to graph databases what e.g., RapidMiner is to SQL databases. It has been used successfully by Lynx Analytics in the past years and now you can use it, too!
Listen to this talk for an intro!
One of the important sources of Lynx’s success is LynxKite. LynxKite is a graph data science platform. It is to graph databases what e.g., RapidMiner is to SQL databases. It has been used successfully by Lynx Analytics in the past years and now you can use it, too!
Listen to this talk for an intro!
András Németh
Andras is the CTO of Lynx Analytics, a graph analytics consultancy. He is in charge of the development of LynxKite, the company’s graph data science platform. He has alsooverseen the technology aspects of various graph projects during his tenure at Lynx. Before Lynx, he worked at Google on ad targeting at YouTube and then on semantic understanding of web pages based on the Knowledge Graph.
Lynx Analytics is a pioneer in driving business value from graph based insights. Our international clientele range from large telcos through banks to transport agencies and governments. Our projects vary from graph based advanced demography estimation, customer satisfaction monitoring through viral campaign design to fiber network and ATM location optimization. The common denominator for all these projects is graph: we always beat state of the art by somehow using the extra information available in some form of relationship data.
One of the important sources of Lynx’s success is LynxKite. LynxKite is a graph data science platform. It is to graph databases what e.g., RapidMiner is to SQL databases. It has been used successfully by Lynx Analytics in the past years and now you can use it, too!
Listen to this talk for an intro!
One of the important sources of Lynx’s success is LynxKite. LynxKite is a graph data science platform. It is to graph databases what e.g., RapidMiner is to SQL databases. It has been used successfully by Lynx Analytics in the past years and now you can use it, too!
Listen to this talk for an intro!
András Vicsek
Andras Vicsek is a researcher, a management consultant and a trainer specialized in organizational and social network analysis with a solid background in change management, team, and organizational development. Before founding Maven7 with his partners, he gained vast experience as the head of several research projects regarding work stress, motivation, loyalty, and employer branding. Andras is the key technical expert of Maven7’s products, services, and related certification training programs Maven7 is providing to its partners.
This pandemic has significantly challenged your organization, teams, work culture and workflows. Trust plays a key role in how your organization adapts to the “new normal” and how the crisis will affect the engagement and productivity of your people. Corporations are under pressure to innovate and become more agile to react to disruption in their industry and changing customer needs as well. This requires a change in the mindset of people, processes and the culture of the organization. Learn how Organizational Network Analysis can help to navigate the uncertainties and nurture a productive and constructive organization.
Milán Janosov
Milan Janosov earned his PhD in Network and Data Science at the Central European University in 2020. His background is in Physics and Complex Systems. During his PhD years, he has also been visiting the BarabasiLab in Boston and the Bell Labs in Cambridge. He is currently the Chief Data Scientist of the data-driven location intelligence start-up, Datapolis. In 2020, he was selected into the Forbes 30U30 list of Hungary.
How can we trace down the hidden role of networks and quantify the presence of luck in success across various creative fields, such as film and music? To answer these questions, in our recent work we put together the tools of network and data science with large-scale datasets to understand i) the timing of the biggest hit during creative careers, ii) the interplay of career- and network peaks over time, iii) and the role of mentorship particularly focusing on star DJs. In my talk, I will give a short overview of these findings with additional notes on its current and potential applications in real-life problems.
Trey Grainger
Trey Grainger is CTO at Presearch.io, the decentralized web search engine, and is the Founder of Searchkernel, a startup helping clients build next-generation, intelligent search applications. He is the author of the books AI-Powered Search and Solr in Action, and is the former Chief Algorithms Officer and SVP of Engineering at Lucidworks, a leading AI-powered search company. He studied information retrieval and web search at Stanford University, received his Masters in Management of Technology from Georgia Tech, and received his Bachelors degree from Furman University in Computer Science, Business, and Philosophy.
Did you know that the text in your content can be directly traversed as a “semantic knowledge graph” to surface the nuanced meaning of both your documents and your users’ queries? When a user types a query into a search box, one of the hardest but most important tasks of an application is to correctly interpret the user’s intent and to return the best answers. Language is fuzzy, though, and the meaning of the words in a query or document may have very different interpretations based upon the domain and user-specific context (for example a “driver” could be a car driver, a kind of golf club, software powering a device, or a motivating force). Understanding the nuanced meaning of both your documents and your users’ queries is important for returning the best matches, but thankfully doing so doesn’t require building out sophisticated machine learning models! In this talk, we’ll introduce you to the concept of “semantic knowledge graphs” and will walk through some fun code examples for leveraging the language of both your users and your content to generate a nuanced understanding of their meaning.
partners
