Semantics-driven attentive few-shot learning over clean and noisy samples

Download

index.pdf

Date

2022-11-01

Author

Baran, Orhun Buğra
Cinbiş, Ramazan Gökberk

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

147
views

0
downloads

Over the last couple of years, few-shot learning (FSL) has attracted significant attention towards minimiz-ing the dependency on labeled training examples. An inherent difficulty in FSL is handling ambiguities resulting from having too few training samples per class. To tackle this fundamental challenge in FSL, we aim to train meta-learner models that can leverage prior semantic knowledge about novel classes to guide the classifier synthesis process. In particular, we propose semantically-conditioned feature attention and sample attention mechanisms that estimate the importance of representation dimensions and training instances. We also study the problem of sample noise in FSL, towards utilizing meta-learners in more realistic and imperfect settings. Our experimental results demonstrate the effectiveness of the proposed semantic FSL model with and without sample noise.(c) 2022 Elsevier B.V. All rights reserved.

Subject Keywords

Few -shot learning, Vision and language integration

URI

https://hdl.handle.net/11511/101119

Journal

NEUROCOMPUTING

DOI

https://doi.org/10.1016/j.neucom.2022.09.121

Collections

Department of Computer Engineering, Article

Suggestions

OpenMETU
Core

Attention mechanisms for semantic few-shot learning Baran, Orhun Buğra; Cinbiş, Ramazan Gökberk; İkizler-Cinbiş, Nazlı; Department of Computer Engineering (2021-9-1) One of the fundamental difficulties in contemporary supervised learning approaches is the dependency on labelled examples. Most state-of-the-art deep architectures, in particular, tend to perform poorly in the absence of large-scale annotated training sets. In many practical problems, however, it is not feasible to construct sufficiently large training sets, especially in problems involving sensitive information or consisting of a large set of fine-grained classes. One of the main topics in machine learning...
Construct Validity of Language Achievement Causal Attribution Scale (LACAS) Erten, İsmail Hakkı; Çağatay, Sibel (2020-01-01) Causal attribution theory has attracted the attention of the scholars recently as it has a pivotal role in learners' motivated behaviour and effort for their future learning experiences. According to this theory, learners' perceived cause of their academic performances can be based on locus of control, controllability or stability. In this way, it might be easier for scholars or teachers to identify the "motivationally-at risk" students and take necessary actions. Although learners might state a wide range ...
THE EFFECTS OF PROCESS ORIENTED GUIDED INQUIRY LEARNING ENVIRONMENT ON STUDENTS' SELF-REGULATED LEARNING SKILLS Şen, Şenol; Yılmaz, Ayhan; Geban, Ömer (2015-01-01) Teaching and learning trend has shifted gradually from a teacher-centered approach to a student-centered approach. Many educational researchers have suggested that teachers must use student-centered learning approaches, such as Process Oriented Guided Inquiry Learning (POGIL). POGIL is an instructional approach combining guided inquiry and cooperative learning in which students are involved in the learning process. The purpose of this study was to investigate the effect of POGIL method compared to tradition...
Closed-form sample probing for training generative models in zero-shot learning Çetin, Samet; Cinbiş, Ramazan Gökberk; Department of Computer Engineering (2022-2-10) Generative modeling based approaches have led to significant advances in generalized zero-shot learning over the past few-years. These approaches typically aim to learn a conditional generator that synthesizes training samples of classes conditioned on class embeddings, such as attribute based class definitions. The final zero-shot learning model can then be obtained by training a supervised classification model over the real and/or synthesized training samples of seen and unseen classes, combined. Therefor...
Pre-Service Physics and Chemistry Teachers' Conceptual Integration of Physics and Chemistry Concepts Tuysuz, Mustafa; BEKTAŞ, Oktay; Geban, Ömer; Öztürk, Gökhan; Yalvac, Bugrahan (2016-06-01) This study examines the pre-service teachers' opinions about conceptual integration (CI) and their understanding of it. A qualitative phenomenology design was used in the study. Data was collected through in-depth semi-structured interviews comprising ten guiding questions. Three pre-service physics and three pre-service chemistry teachers participated conveniently in the study. While the pre-service chemistry teachers took a course regarding the CI explicitly taught, the pre-service physics teachers were n...

Citation Formats

O. B. Baran and R. G. Cinbiş, “Semantics-driven attentive few-shot learning over clean and noisy samples,” NEUROCOMPUTING, vol. 513, pp. 59–69, 2022, Accessed: 00, 2022. [Online]. Available: https://hdl.handle.net/11511/101119.