Cross-stitched Multi-modal Encoders
원문 초록
In this paper, we propose a novel architecture for multi-modal speech and text input. We combine pretrained speech and text encoders using multi-headed cross-modal attention and jointly fine-tune on the target problem. The resultant architecture can be used for continuous token-level classification or utterance-level prediction acting on simultaneous text and speech. The resultant encoder efficiently captures both acoustic-prosodic and lexical information. We compare the benefits of multi-headed attention-based fusion for multi-modal utterance-level classification against a simple concatenation of pre-pooled, modality-specific representations. Our model architecture is compact, resource efficient, and can be trained on a single consumer GPU card.
함께 살펴볼 논문
공통 태그를 근거로 연결했습니다. 실제 인용 관계나 논문의 우수성을 뜻하지 않습니다.
Arctic Inference with Shift Parallelism: Fast and Efficient Open Source Inference System for Enterprise AI
Efficient Calisthenics Skills Classification through Foreground Instance Selection and Depth Estimation
Relative Entropy Pathwise Policy Optimization
DEARLi: Decoupled Enhancement of Recognition and Localization for Semi-supervised Panoptic Segmentation
AI-Enhanced Pediatric Pneumonia Detection: A CNN-Based Approach Using Data Augmentation and Generative Adversarial Networks (GANs)
Scaling Attention to Very Long Sequences in Linear Time with Wavelet-Enhanced Random Spectral Attention (WERSA)
토론
이 논문과 연결된 포럼 글이 아직 없습니다.