Papers with Learning representations
Strong and Simple Baselines for Multimodal Utterance Embeddings (N19-1)
Copied to clipboard
| Challenge: | Human language is a rich multimodal signal consisting of spoken words, facial expressions, body gestures, and vocal intonations. |
| Approach: | They propose two simple but strong baselines to learn embeddings of multimodal utterances by factorizing the utterant into unimodal factors. |
| Outcome: | The proposed models show that they can be derived in closed form while maintaining simplicity and efficiency during learning and inference. |
Learning Transferable Feature Representations Using Neural Networks (P19-1)
Copied to clipboard
| Challenge: | Traditional domain adaptation algorithms learn common representations which suffer from transfer loss when the source specific characteristics detract their ability to represent the target data. |
| Approach: | They propose to segregate source specific representation from the common representation and use it to learn a two-part representation which captures source specific characteristics while the second part captures the truly common representation. |
| Outcome: | The proposed representation outperforms existing learning algorithms on the source learning as well as cross-domain tasks on multiple datasets. |