Anna Fensel

Dominik Filipiak, Andrzej Zapała, Piotr Tempczyk et al.

We present Polite Teacher, a simple yet effective method for the task of semi-supervised instance segmentation. The proposed architecture relies on the Teacher-Student mutual learning framework. To filter out noisy pseudo-labels, we use confidence thresholding for bounding boxes and mask scoring for masks. The approach has been tested with CenterMask, a single-stage anchor-free detector. Tested on the COCO 2017 val dataset, our architecture significantly (approx. +8 pp. in mask AP) outperforms the baseline at different supervision regimes. To the best of our knowledge, this is one of the first works tackling the problem of semi-supervised instance segmentation and the first one devoted to an anchor-free detector.

Dominik Filipiak, Anna Fensel, Agata Filipowska

We propose KGTN-ens, a framework extending the recent Knowledge Graph Transfer Network (KGTN) in order to incorporate multiple knowledge graph embeddings at a small cost. We evaluate it with different combinations of embeddings in a few-shot image classification task. We also construct a new knowledge source - Wikidata embeddings - and evaluate it with KGTN and KGTN-ens. Our approach outperforms KGTN in terms of the top-5 accuracy on the ImageNet-FS dataset for the majority of tested settings.

Elwin Huaman, Amar Tauqeer, Anna Fensel

The performance of applications, such as personal assistants and search engines, relies on high-quality knowledge bases, a.k.a. Knowledge Graphs (KGs). To ensure their quality one important task is knowledge validation, which measures the degree to which statements or triples of KGs are semantically correct. KGs inevitably contain incorrect and incomplete statements, which may hinder their adoption in business applications as they are not trustworthy. In this paper, we propose and implement a Validator that computes a confidence score for every triple and instance in KGs. The computed score is based on finding the same instances across different weighted knowledge sources and comparing their features. We evaluate our approach by comparing its results against a baseline validation. Our results suggest that we can validate KGs with an f-measure of at least 75%. Time-wise, the Validator, performed a validation of 2530 instances in 15 minutes approximately. Furthermore, we give insights and directions toward a better architecture to tackle KG validation.