Refine
Year of publication
Document Type
- Article (687)
- Book (221)
- Part of a Book (70)
- Conference Proceeding (68)
- Other (50)
- Review (13)
- Working Paper (8)
- Contribution to a Periodical (5)
- Doctoral Thesis (5)
- Lecture (5)
- Report (4)
- Diploma Thesis (1)
- Talk (1)
Language
- German (924)
- English (207)
- Portuguese (6)
- Spanish (1)
Keywords
- Führung (8)
- Leadership (8)
- Telekommunikationsmarkt (6)
- Datenschutz (5)
- Motivation (5)
- Betriebswirtschaftslehre (4)
- Finanzierung (4)
- Liquiditätspolitik (4)
- Datenschutzgrundverordnung (3)
- Self-Leadership (3)
- Coaching (2)
- Datenschutzrecht (2)
- EU-Datenschutzgrundverordnung (2)
- Internationales Recht / Europarecht (2)
- Prozessautomatisierung (2)
- Prozessverbesserung (2)
- Robotic Process Automation (2)
- Selektionskriterien (2)
- regulation (2)
- 3. EU Legislativpaket (1)
Institute
- Fachbereich Wirtschaftswissenschaften (1138) (remove)
Allgemeines Steuerrecht
(2015)
Supervised machine learning and deep learning require a large amount of labeled data, which data scientists obtain in a manual, and time-consuming annotation process. To mitigate this challenge, Active Learning (AL) proposes promising data points to annotators they annotate next instead of a subsequent or random sample. This method is supposed to save annotation effort while maintaining model performance.
However, practitioners face many AL strategies for different tasks and need an empirical basis to choose between them. Surveys categorize AL strategies into taxonomies without performance indications. Presentations of novel AL strategies compare the performance to a small subset of strategies. Our contribution addresses the empirical basis by introducing a reproducible active learning evaluation (ALE) framework for the comparative evaluation of AL strategies in NLP.
The framework allows the implementation of AL strategies with low effort and a fair data-driven comparison through defining and tracking experiment parameters (e.g., initial dataset size, number of data points per query step, and the budget). ALE helps practitioners to make more informed decisions, and researchers can focus on developing new, effective AL strategies and deriving best practices for specific use cases. With best practices, practitioners can lower their annotation costs. We present a case study to illustrate how to use the framework.
AI-based systems are nearing ubiquity not only in everyday low-stakes activities but also in medical procedures. To protect patients and physicians alike, explainability requirements have been proposed for the operation of AI-based decision support systems (AI-DSS), which adds hurdles to the productive use of AI in clinical contexts. This raises two questions: Who decides these requirements? And how should access to AI-DSS be provided to communities that reject these standards (particularly when such communities are expert-scarce)? This chapter investigates a dilemma that emerges from the implementation of global AI governance. While rejecting global AI governance limits the ability to help communities in need, global AI governance risks undermining and subjecting health-insecure communities to the force of the neo-colonial world order. For this, this chapter first surveys the current landscape of AI governance and introduces the approach of relational egalitarianism as key to (global health) justice. To discuss the two horns of the referred dilemma, the core power imbalances faced by health-insecure collectives (HICs) are examined. The chapter argues that only strong demands of a dual strategy towards health-secure collectives can both remedy the immediate needs of HICs and enable them to become healthcare independent.