Unifying Visual-Semantic Embeddings with Multimodal Neural Language Models
Ryan Kiros, Ruslan Salakhutdinov, Richard S. Zemel University of Toronto Canadian Institute for Advanced Research {rkiros, rsalakhu,
Abstract
Inspired by recent advances in multimodal learning and machine translation, we introduce an encoder-decoder pipeline that learns (a): a multimodal joint embedding space with images and text and (b): a novel language model for decoding distributed representations from our space. Our pipeline effectively unifies joint image-text embedding models with multimodal neural language models. We introduce the structure-content neural language model that disentangles the structure of a sentence to its content, conditioned on representations produced by the encoder. The encoder allows one to rank images and sentences while the decoder can generate novel descriptions from scratch. Using LSTM to encode sentences, we match the state-of-the-art performance on Flickr8K and Flickr30K without using object detections. We also set new best results when using the 19-layer Oxford convolutional network. Furthermore we show that with linear encoders, the learned embedding space captures multimodal regularities in terms of vector space arithmetic e.g. *image of a blue car* - "blue" + "red" is near images of red cars. Sample captions generated for 800 images are made available for comparison.
中文速览
图像描述生成领域长期面临一个难题：如何让模型既能判断图文是否匹配（排序），又能从头生成流畅自然的图像描述（生成）。作者提出了一套编码器-解码器流水线，编码器用LSTM将句子与图像映射到同一个多模态向量空间，通过成对排序损失让图文表示相互靠近；解码器则是一种新型语言模型——结构-内容神经语言模型（Structure-Content Neural Language Model，SC-NLM），它将句子的句法结构与语义内容解耦，以编码器产生的多模态表示为条件生成新描述。实验表明，该方法在Flickr8K和Flickr30K上达到当时最优的图文排序性能，配合19层牛津卷积网络还刷新了最佳成绩；此外，学到的向量空间还支持类似词向量的多模态类比运算，例如"蓝色汽车的图像向量 − '蓝色' + '红色'"所指向的位置恰好接近红色汽车的图像。这项工作的意义在于，它首次将图文联合嵌入模型与多模态神经语言模型统一在同一框架下，为后续图像描述生成与视觉问答研究奠定了方法论基础。
原文 arXiv:1411.2539;中英对照 + 大白话阅读 https://aha.fim.ai/paper/1411.2539v1