Alleviating Sequence Information Loss with Data Overlapping and Prime Batch Sizes
In sequence modeling tasks the token order matters, but this information can be partially lost due to the discretization of the sequence into data points. In this paper, we study the imbalance between the way certain token pairs are included in data points and others are not. We denote this a token order imbalance (TOI) and we link the partial sequence information loss to a diminished performance of the system as a whole, both in text and speech processing tasks. We then provide a mechanism to leverage the full token order information -Alleviated TOI- by iteratively overlapping the token composition of data points. For recurrent networks, we use prime numbers for the batch size to avoid redundancies when building batches from overlapped data points. The proposed method achieved state of the art performance in both text and speech related tasks.
Code (1)
Tasks
Language ModellingSimilar Papers 제목 키워드 기반
Encoder-decoder Network as Loss Function for Summarization
We present a new approach to defining a sequence loss function to train a summarizer by using a secondary encoder-decoder as a loss function, alleviating a shortcoming of word level training for sequence outputs. The tec…
DecoderNews SummarizationBLADE: Box-Level Supervised Amodal Segmentation through Directed Expansion
Perceiving the complete shape of occluded objects is essential for human and machine intelligence. While the amodal segmentation task is to predict the complete mask of partially occluded objects, it is time-consuming an…
Amodal Instance SegmentationSegmentationOAMatcher: An Overlapping Areas-based Network for Accurate Local Feature Matching
Local feature matching is an essential component in many visual applications. In this work, we propose OAMatcher, a Tranformer-based detector-free method that imitates humans behavior to generate dense and accurate match…
Cross-domain Transfer of Valence Preferences via a Meta-optimization Approach
Cross-domain recommendation offers a potential avenue for alleviating data sparsity and cold-start problems. Embedding and mapping, as a classic cross-domain research genre, aims to identify a common mapping function to …
Meta-LearningSelf-Supervised LearningUnifying Isolated and Overlapping Audio Event Detection with Multi-Label Multi-Task Convolutional Recurrent Neural Networks
We propose a multi-label multi-task framework based on a convolutional recurrent neural network to unify detection of isolated and overlapping audio events. The framework leverages the power of convolutional recurrent ne…
Event DetectionMulti-Label ClassificationMUlTI-LABEL-ClASSIFICATION