PhD on Generative AI for Insight Retrieval from Structured Data

Goal of this PhD project

Despite the abundance of structured data, many important insights from this kind of data are left unnoticed as they are difficult to surface or organizations do not have sufficient expertise to extract them. This project will explore emerging paradigms such as table representation learning (TRL), retrieval- augmented generation, generative retrieval, agentic systems, and conversational user interfaces, to automatically surface relevant insights from relational databases and data lakes, in a robust and efficient manner. This PhD project is supervised by Dr. Madelon Hulsebos (CWI) and Prof. dr. Maarten de Rijke (UvA).

What you will be doing

• Develop and execute a 4-year research agenda around insight retrieval from structured data.
• Actively collaborate with other researchers in the DataLibra project (students, PhDs, postdocs,PIs) and external collaborators.
• Communicate research outcomes through papers and presentations at conferences, workshops and other (scientific) gatherings.
• Assist in relevant teaching activities at universities, such as thesis supervision and assisting in courses.

Informatiespecialist Bibliotheek (x/v/m) voor 32 - 40 uur per week

De bibliotheek van het Centrum Wiskunde en Informatica (CWI) heeft een unieke collectie boeken en rapporten over wiskunde en informatica die toegankelijk is voor wetenschappers en andere belangstellenden van binnen en buiten CWI. Kom jij binnenkort onze papieren en digitale informatievoorziening beheren en onze wetenschappers wegwijs maken in de bibliotheek?

Functiebeschrijving
Als informatiespecialist voor de CWI Bibliotheek beheer je de analoge en digitale collecties en help je onderzoekers en promovendi bij het vinden van wetenschappelijk werk. Je organiseert de open opstellingen en archieven en brengt bijzondere materialen onder de aandacht bij onze bezoekers. Als je kansen ziet om de informatievoorziening te verbeteren dan bieden wij je alle ruimte om dit op te pakken.

Het is een uitdagende, zelfstandige functie in een klein team van vier collega’s die samen meer dan 250 wetenschappers van binnen en buiten CWI ondersteunen.

In de rol van informatiespecialist:

• Beheer je de modules binnen het WMS bibliotheeksysteem van OCLC:

recordmanager, acquisitions, licenties, uitleen en interbibliothecair leenverkeer

(IBL)

• Help je wetenschappers, bezoekers en externen om literatuur te vinden
• Je houdt de collectie up-to-date (op papier en digitaal), catalogiseert en samen

met de coördinator werk je aan gebruikersonderzoeken

• Je houdt je op de hoogte van nieuwe ontwikkelingen in het veld en neemt deel

aan de WMS gebruikersgroep

• Je beheert de archieven en digitaliseert stukken op aanvraag en naar eigen

inzicht

• Je beheert informatie in andere digitale systemen, bijvoorbeeld de foto

database en dissertaties in de Institutional Repository

• Wanneer nodig help je mee bij andere werkzaamheden binnen het team,

bijvoorbeeld bij data invoer in de Institutional Repository

Postdocs, ERC Advanced Grant Project FLEX (m/v/x)

Most statistical methods require that all aspects of data collection and inference are determined in advance, independently of the data. These include when to stop collecting data, what decisions can be made (i.e. accept/reject null hypothesis), and how to measure their quality (e.g. loss function/significance level). This is wildly at odds with the flexibility required in practice! It is a leading cause of the replicability crisis in the applied sciences.

The project aims to develop novel theory and methods of statistics in which all data-collection and decision-aspects may be unknown in advance, possibly imposed post-hoc. Yet the new theory will provide valid small-sample error control and uncertainty quantification. The theory will be based on far-reaching extensions of e-values, e-processes, and the e-posterior, flexible alternatives for p-values and Bayes factors that have been developed over the last five years. Other key topics include anytime-valid confidence sequences, concentration inequalities, martingales and the foundations of statistics (likelihood principle, complete class theorems).

The vacancies are for two three-year postdocs to do research within the FLEX project.

PhD students, ERC Advanced Grant Project FLEX (m/v/x)

Most statistical methods require that all aspects of data collection and inference are determined in advance, independently of the data. These include when to stop collecting data, what decisions can be made (i.e. accept/reject null hypothesis), and how to measure their quality (e.g. loss function/significance level). This is wildly at odds with the flexibility required in practice! It is a leading cause of the replicability crisis in the applied sciences.

The FLEX project aims to develop novel theory and methods of statistics in which all data-collection and decision-aspects may be unknown in advance, possibly imposed post-hoc. Yet this new theory will provide small-sample frequentist error control and uncertainty quantification. The theory will be based on far-reaching extensions of e-values, e-processes, and the e-posterior, flexible alternatives for p-values and Bayes factors that have been developed over the last five years. Other key topics include anytime-valid confidence sequences and the foundations of statistics (likelihood principle, complete class theorems).

peter

The vacancies are for two four-year PhD students to do research within the FLEX project. The students will join the Machine Learning group at CWI, and receive their PhD degree from Leiden University.

Director Centrum Wiskunde & Informatica (CWI) (m/v/x) 0,8 - 1,0 fte

What will you do?

• As director you are responsible for steering the institute, focusing on the overall strategic goals, leading the development of strategic plans and overseeing their implementation.
• You represent the institute, and maintain relationships with Dutch universities and national, European and international partners in science, industry and society.
• Together with the institute manager, you account for CWI’s results and plans with the board of NWO-I, twice a year.

In more detail:

• You consolidate CWI’s functioning as a national centre, with well-developed collaborative ties to Dutch universities and you contribute to national research agendas and initiatives.
• You advise the board of NWO on CWI’s scientific mission and strategy.
• You effectively promote the interests of CWI and Dutch science in the area of mathematics and computer science.
• You constructively align CWI’s research agenda with its stakeholders, e.g., universities and partners in strategic alliances from public and private organisations.
• You develop policies to establish a diverse and inclusive work environment.
• You ensure a stimulating work environment.
• You are responsible for the institute’s staff and talent policies.

How does CWI’s organizational structure assist you?

• Together with the institute manager (responsible for finance & operations) you are CWI’s directorate. Together with four senior researchers, the directorate forms CWI’s management team, which meets every other week.
• CWI’s institute manager, policy advisor, manager communications, manager HR and several senior scientists with strategic roles report directly to you.
• You regularly meet with CWI’s fifteen research group leaders, with the works council, and with support staff managers. Twice a year you consult with CWI’s advisory bodies, being the national Institute Advisory Council and the international Scientific Advisory Council.