1 |
Comic Spin: A Comic Creation Tool Enabling Self-Expression for People with Aphasia
|
|
|
|
BASE
|
|
Show details
|
|
2 |
Modelling Group Dynamics with SYMLOG and Snowdrift for Intelligent Classroom Environment
|
|
|
|
BASE
|
|
Show details
|
|
5 |
“Just Not Together”: The Experience of Videoconferencing for People with Aphasia during the Covid-19 Pandemic
|
|
|
|
BASE
|
|
Show details
|
|
8 |
A Framework for Quality Assessment of Semantic Annotations of Tabular Data
|
|
|
|
BASE
|
|
Show details
|
|
9 |
Delivering group support for people with aphasia in a virtual world: experiences of service providers
|
|
|
|
BASE
|
|
Show details
|
|
10 |
Anti-transfer learning for task invariance in convolutional neural networks for speech processing
|
|
|
|
Abstract:
We introduce the novel concept of anti-transfer learning for speech processing with convolutional neural networks. While transfer learning assumes that the learning process for a target task will benefit from re-using representations learned for another task, anti-transfer avoids the learning of representations that have been learned for an orthogonal task, i.e., one that is not relevant and potentially confounding for the target task, such as speaker identity for speech recognition or speech content for emotion recognition. This extends the potential use of pre-trained models that have become increasingly available. In anti-transfer learning, we penalize similarity between activations of a network being trained on a target task and another one previously trained on an orthogonal task, which yields more suitable representations. This leads to better generalization and provides a degree of control over correlations that are spurious or undesirable, e.g. to avoid social bias. We have implemented anti-transfer for convolutional neural networks in different configurations with several similarity metrics and aggregation functions, which we evaluate and analyze with several speech and audio tasks and settings, using six datasets. We show that anti-transfer actually leads to the intended invariance to the orthogonal task and to more appropriate features for the target task at hand. Anti-transfer learning consistently improves classification accuracy in all test cases. While anti-transfer creates computation and memory cost at training time, there is relatively little computation cost when using pre-trained models for orthogonal tasks. Anti-transfer is widely applicable and particularly useful where a specific invariance is desirable or where labeled data for orthogonal tasks are difficult to obtain on a given dataset but pre-trained models are available.
|
|
Keyword:
P Philology. Linguistics; QA75 Electronic computers. Computer science; RC0321 Neuroscience. Biological psychiatry. Neuropsychiatry
|
|
URL: https://doi.org/10.1016/j.neunet.2021.05.012 https://openaccess.city.ac.uk/id/eprint/26378/1/2006.06494v2.pdf https://openaccess.city.ac.uk/id/eprint/26378/
|
|
BASE
|
|
Hide details
|
|
11 |
Comparing the environmental impacts of recipes from four different recipe databases using Natural Language Processing
|
|
|
|
BASE
|
|
Show details
|
|
12 |
A World Full of Stereotypes? Further Investigation on Origin and Gender Bias in Multi-Lingual Word Embeddings ...
|
|
|
|
BASE
|
|
Show details
|
|
13 |
Calculating the optimal step of arc-eager parsing for non-projective trees
|
|
|
|
BASE
|
|
Show details
|
|
14 |
Few-shot linguistic grounding of visual attributes and relations using gaussian kernels
|
|
|
|
BASE
|
|
Show details
|
|
15 |
Automated detection of Hainan gibbon calls for passive acoustic monitoring
|
|
|
|
BASE
|
|
Show details
|
|
16 |
Exploring the characteristics of abusive behaviour in online social media settings
|
|
|
|
BASE
|
|
Show details
|
|
17 |
Deep Scattering and End-to-End Speech Models towards Low Resource Speech Recognition
|
|
|
|
BASE
|
|
Show details
|
|
18 |
Framework for Composition of Domain Specific Languages and the Effect of Composition on Re-use of Translation Rules
|
|
|
|
BASE
|
|
Show details
|
|
19 |
Superstar to Superhuman: Scarlett Johansson, an ‘Ideal’ Embodiment of the Posthuman Female in Science Fiction and Media?
|
|
|
|
BASE
|
|
Show details
|
|
20 |
Qualitative-geometric ‘surrounds’ relations between disjoint regions
|
|
|
|
BASE
|
|
Show details
|
|
|
|