Data Value: Assessment and Evolution

Description

Data and data­-driven artificial intelligence are impacting virtually every aspect of society including commerce, science, medicine, government, finance and education. Data drives value in all these various domains, but we know surprisingly little about how value accrues to data as it progresses through its lifecycle — collection, wrangling and integration, modeling and analysis, decision making, and curation.

Data value arises through its ability to foster new products and business models and to enable new discoveries across science, engineering and humanities. However, as a society we are also starting to become aware of the many harms data can bring and their associated risks. When used to inform decisions that affect individuals, for example, data can perpetuate or emphasize existing biases. Combining data from disparate sources can enable previously unseen insights but can also expose private information, either intentionally or unintentionally. Untrustworthy data can also wreak havoc on society, negatively impacting individual lives or putting democracy at risk.

Thus, a challenge we face today is to design systems and data-­driven organizations that maximize data’s positive impact while minimizing the negative effects. But we cannot do this without understanding how data contributes to both good and bad outcomes: the crux of the problem is to understand what is the value of data and how does that value change over time, through various processing steps, and when being used in changing contexts.

The aim of this 3­-day workshop is to explore these questions about data value and to discuss approaches to answering them. We will approach the questions from different angles: economic theory, statistics, data semantics, privacy, data markets and software platforms, to name a few. We intend to produce a report that incorporates the outcomes of the workshop.

In addition to the talks, we will hold:

• sessions with contributed talks by students and postdocs, and
• a panel with companies that are building data markets.

Schedule

Videos

Information Sale with Externalities

Munther Dahleh
June 6, 2022

Data, Competition, and Digital Platforms

Dirk Bergemann
June 6, 2022

Optimal Advertising for Information Products

Yiling Chen
June 6, 2022

Privacy Guarantees and Inference with an Oblivious Data Curator

Denis Nekipelov
June 6, 2022

Quantifying Information and Uncertainty

Emir Kamenica
June 6, 2022

“Soul-Bound” Tokens, PETs, and More: Speculative Data Governance Structures

Matt Prewitt
June 7, 2022

The inherent instability of top-down valuation methods: bottom-up data trusts and their political, economic and social potential

Sylvie Delacroix
June 7, 2022

Data Sharing starts at home

Juan Sequeda
June 7, 2022

Progressions of data Shapley valuation and its applications

James Zou
June 7, 2022

Industry Panel

Industry Panel
June 7, 2022

Optimal Pricing of Information

Haifeng Xu
June 8, 2022

Query-based and Model-based Data Pricing

Paris Koutris
June 8, 2022

Value of Data for ML, and some PTIME Valuation Results

Ce Zhang
June 8, 2022

The Long Tail of Data Curation and its Impact of the Value of Data

Robert Grossman
June 8, 2022

Data, the Fundamental Particle of Economics

Nicole Immorlica
June 8, 2022

Fair and Efficient Valuation in Data Marketplaces

Jian Pei
June 8, 2022