Polarization on Social Media Aristides Gionis

Which topics spark the most heated debates on social media? How do social media users interact when they disagree? Can we nudge them towards constructive disagreement? Answering these questions helps us understand large-scale societal mechanisms - including phenomena such as partisan sharing, echo chambers, and filter bubbles, which might be detrimental to the democratic process. It also allows us to develop tools and techniques that can be used by journalists to understand which issues divide the public; social scientists to understand how polarization is manifested in social interactions; and web users to improve their experience online. This tutorial presents a systematic review of polarization as manifested online, and in particular on social media. In the beginning, we clarify the concept of polarization and the related nomenclature, drawing from the social and political sciences. Subse- quently, we review algorithmic techniques for detection, quantification, and mitigation of polarization. We conclude the tutorial by presenting open challenges and promising research directions for Web Scientists.

Slides available (PDF)

Coffee break

Polarization on Social Media Aristides Gionis

Which topics spark the most heated debates on social media? How do social media users interact when they disagree? Can we nudge them towards constructive disagreement? Answering these questions helps us understand large-scale societal mechanisms - including phenomena such as partisan sharing, echo chambers, and filter bubbles, which might be detrimental to the democratic process. It also allows us to develop tools and techniques that can be used by journalists to understand which issues divide the public; social scientists to understand how polarization is manifested in social interactions; and web users to improve their experience online. This tutorial presents a systematic review of polarization as manifested online, and in particular on social media. In the beginning, we clarify the concept of polarization and the related nomenclature, drawing from the social and political sciences. Subse- quently, we review algorithmic techniques for detection, quantification, and mitigation of polarization. We conclude the tutorial by presenting open challenges and promising research directions for Web Scientists.

Slides available (PDF)

Lunch

Misinformation and Information DisorderKalina Bontcheva

This tutorial introduces a theoretical framework for studying the information disorder. It then presents the latest studies and state-of-the-art approaches for analysing online disinformation campaigns, detecting online abuse, and quantifying their impact on the target audiences. The material is structured under six key questions: What is being spread? Who is spreading it? When it spreads? Where it spreads? Why it spreads? How it spreads? It concludes discussing outstanding challenges and hot topics for future work.

Slides available (PDF)

Coffee break

Misinformation and Information Disorder Kalina Bontcheva

This tutorial introduces a theoretical framework for studying the information disorder. It then presents the latest studies and state-of-the-art approaches for analysing online disinformation campaigns, detecting online abuse, and quantifying their impact on the target audiences. The material is structured under six key questions: What is being spread? Who is spreading it? When it spreads? Where it spreads? Why it spreads? How it spreads? It concludes discussing outstanding challenges and hot topics for future work.

Slides available (PDF)

The Frontiers of Machine Learning John Shawe-Taylor

Slides available (PDF)

Coffee break

The Frontiers of Machine Learning John Shawe-Taylor

Slides available (PDF)

Lunch

Social and ethical implications of AI Nello Cristianini

We have deployed a version of Artificial Intelligence at the centre of the global data infrastructure, in a way that cannot be avoided. Consequential decisions are constantly being made by AI agents about us, some minor such as recommending a video, some major, such as mortgage or health or even judiciary decisions. Many workers now are coordinated by intelligent algorithms, for example in the delivery and care-sharing economies. What are the consequences for society and the individual, of decisions that are often not explained, cannot be easily questioned, and have been suspected of being biased? As we cannot avoid this technology, we need to develop ways to live with it. Maximising accuracy or predictive performance is no longer enough, when the object of these decisions are humans. A new set of considerations is becoming important in the design of AI systems. Fixing these problem might be difficult, if we consider that some of the very shortcuts that gave us a cheap path to AI are implicated in some of the main problems that we are observing.

Slides available (PDF)

Coffee break

Trip to the beach

Social Dinner at InVilla, via Litoranea 18, Marina di Pisa

Algorithm and Mechanism Design for Online Markets Stefano Leonardi

Online markets currently form an important share of the global economy. The Internet hosts classical markets (real-estate, stocks, e-commerce) as well allowing new data driven markets with previously unknown features (web-based advertisement, viral marketing, digital goods, crowdsourcing, sharing economy). Algorithms play a central role in the design of economic mechanisms involved in online markets. For example, algorithms electronic auctions, trade stocks, adjusts prices dynamically, and harvest big data to provide economic information. In these lectures we present some of the main challenges of algorithm and mechanism design in online markets. We also present recent results and challenging open problems on two specific research areas: Two-sided Markets and Online Labour Marketpalces.

Slides (1) available (PDF)
Slides (2) available (PDF)

Coffee break

Algorithm and Mechanism Design for Online Markets Stefano Leonardi

Online markets currently form an important share of the global economy. The Internet hosts classical markets (real-estate, stocks, e-commerce) as well allowing new data driven markets with previously unknown features (web-based advertisement, viral marketing, digital goods, crowdsourcing, sharing economy). Algorithms play a central role in the design of economic mechanisms involved in online markets. For example, algorithms electronic auctions, trade stocks, adjusts prices dynamically, and harvest big data to provide economic information. In these lectures we present some of the main challenges of algorithm and mechanism design in online markets. We also present recent results and challenging open problems on two specific research areas: Two-sided Markets and Online Labour Marketpalces.

Slides (1) available (PDF)
Slides (2) available (PDF)

Lunch

Business Process Analytics: from insights to Predictions Marlon Dumas

Business process analytics is a body of methods for analyzing data generated during the execution of business processes, in order to extract insights about weaknesses and improvement opportunities, both at the tactical and operational levels. Tactical process analytics methods (also known as process mining methods) allow us to understand how a given process is executed, if and how its execution deviates with respect to expected or normative pathways, and what factors contribute to poor process performance or undesirable outcomes. Meantime, operational process analytics methods allow us to monitor ongoing executions of a business process in order to predict future states and undesirable outcomes at runtime (predictive process monitoring). Existing methods in this space allow us to predict, for example, which task will be executed next in a case, when, and who will perform it? When will an ongoing case complete? What will its outcome be and how can negative outcomes be avoided? This lecture will present a framework for conceptualizing business process analytics methods and applications. The lecture will provide an overview of state-of-art methods and tools in the field and will outline open challenges and research opportunities, particularly in relation to explainability and actionability of predictions.

Slides available (PDF)

Lunch

Business Process Analytics: from insights to Predictions Marlon Dumas

Business process analytics is a body of methods for analyzing data generated during the execution of business processes, in order to extract insights about weaknesses and improvement opportunities, both at the tactical and operational levels. Tactical process analytics methods (also known as process mining methods) allow us to understand how a given process is executed, if and how its execution deviates with respect to expected or normative pathways, and what factors contribute to poor process performance or undesirable outcomes. Meantime, operational process analytics methods allow us to monitor ongoing executions of a business process in order to predict future states and undesirable outcomes at runtime (predictive process monitoring). Existing methods in this space allow us to predict, for example, which task will be executed next in a case, when, and who will perform it? When will an ongoing case complete? What will its outcome be and how can negative outcomes be avoided? This lecture will present a framework for conceptualizing business process analytics methods and applications. The lecture will provide an overview of state-of-art methods and tools in the field and will outline open challenges and research opportunities, particularly in relation to explainability and actionability of predictions.

Slides available (PDF)

Ethics for data science Jeroen van den Hoven

Coffee break

Ethics for data science Jeroen van den Hoven

Lunch

XAI - Science and technology for the explanation of AI decision making Dino Pedreschi

Black box AI systems for automated decision making, often based on machine learning over (big) data, map a user’s features into a class or a score without exposing the reasons why. This is problematic not only for the lack of transparency, but also for possible biases inherited by the algorithms from human prejudices and collection artifacts hidden in the training data, which may lead to unfair or wrong decisions. The future of AI lies in enabling people to collaborate with machines to solve complex problems. Like any efficient collaboration, this requires good communication, trust, clarity and understanding. Explainable AI addresses such challenges and for years different AI communities have studied such topic, leading to different definitions, evaluation protocols, motivations, and results. This lecture provides a reasoned introduction to the work of Explainable AI (XAI) to date, and surveys the literature with a focus on machine learning and symbolic AI related approaches. We motivate the needs of XAI in real-world and large-scale application, while presenting state-of-the-art techniques and best practices, as well as discussing the many open challenges.

Slides available (PDF)

XAI - Science and technology for the explanation of AI decision making Fosca Giannotti

Black box AI systems for automated decision making, often based on machine learning over (big) data, map a user’s features into a class or a score without exposing the reasons why. This is problematic not only for the lack of transparency, but also for possible biases inherited by the algorithms from human prejudices and collection artifacts hidden in the training data, which may lead to unfair or wrong decisions. The future of AI lies in enabling people to collaborate with machines to solve complex problems. Like any efficient collaboration, this requires good communication, trust, clarity and understanding. Explainable AI addresses such challenges and for years different AI communities have studied such topic, leading to different definitions, evaluation protocols, motivations, and results. This lecture provides a reasoned introduction to the work of Explainable AI (XAI) to date, and surveys the literature with a focus on machine learning and symbolic AI related approaches. We motivate the needs of XAI in real-world and large-scale application, while presenting state-of-the-art techniques and best practices, as well as discussing the many open challenges.

Slides available (PDF)

Coffee break

Informatica50: Pisa e l'Intelligenza Artificiale panel (in Italiano) moderato da Franco Turini e Fosca Giannotti

L'informatica pisana ha contribuito anche alla nascita della ricerca in Intelligenza Artificiale dando vita a molteplici scuole che hanno gemmato anche in molti altri atenei italiani definendo l'agenda di ricerca e l'innovazione della formazione superiore su temi fondazionali quali: i linguaggi per l'IA, la programmazione logica, la rappresentazione della conoscenza, l'elaborazione del linguaggio naturale, l'apprendimento automatico e la scienza dei dati. Intervengono: Franco Sirovich Luigia Carlucci Aiello, Gianfranco Prini, Giuseppe Attardi, Maria Simi, Roberto Barbuti, Dino Pedreschi, Alessio Micheli.