4 papers
When Confidence Takes the Wrong Path: Diagnosing Retrieval-State Lock-In in RAG
Sahib Julka
The trustworthiness of a retrieval-augmented generation (RAG) system depends on more than the answer it returns, yet many black-box uncertainty methods still read agreement among s…
Towards an Improved Metric for Evaluating Disentangled Representations
Sahib Julka, Yashu Wang, Michael Granitzer
Disentangled representation learning plays a pivotal role in making representations controllable, interpretable and transferable. Despite its significance in the domain, the quest…
DriftGAN: Using historical data for Unsupervised Recurring Drift Detection
Christofer Fellicious, Sahib Julka, Lorenz Wendlinger +1
In real-world applications, input data distributions are rarely static over a period of time, a phenomenon known as concept drift. Such concept drifts degrade the model's predictio…
LLMs in the Loop: Leveraging Large Language Model Annotations for Active Learning in Low-Resource Languages
Nataliia Kholodna, Sahib Julka, Mohammad Khodadadi +2
Low-resource languages face significant barriers in AI development due to limited linguistic resources and expertise for data labeling, rendering them rare and costly. The scarcity…