In this paper, we propose a novel method called cross-modal memory fusion network (CMFN) that explicitly learns both modal-specific and cross-modal dynamics for imputing the missing values in multi-modal sequential learning tasks. Experiments on two datasets demonstrate that our method outperforms ...
第一篇将侧重 "multi-modal" 和 "application", 介绍相关概念与研究背景;第二、三篇会侧重"algorithm" 介绍这个方向研究的技术路线,其中第二篇介绍基于 GAN 的追求公共子空间的 cross-modal 检索;第三篇则从modal抽象成更一般的domain,并且将多域扩展到单域,总结分析单/多域匹配问题,主要介绍基于contrastivelearnin...
Previous studies mainly follow the encoder-decoder paradigm and focus on the aspect of text generation, with few studies considering the importance of cross-modal mappings and explicitly exploit such mappings to facilitate radiology report generation. In this paper, we propose a cross-modal memory ...
releases GABAergic inhibition to permit specific microcircuits within KC-spanning serotonergic neurons to function as an excitatory bridge between the previously ‘modality-selective’ KC streams. Cross-modal binding thereby expands the KCs representing the memory engram for each modality into those represen...
Memory-retrieval processing of cross-modal sensory preconditioning is vital for understanding the plasticity underlying the interactions between modalities. As part of the sensory preconditioning paradigm, it has been hypothesized that the conditioned response to an unreinforced cue depends on the memory of...
Two experiments used visual-, verbal-, and haptic-interference tasks during encoding (Experiment 1) and retrieval (Experiment 2) to examine mental representation of familiar and unfamiliar objects in visual/haptic crossmodal memory. Three competing theories are discussed, which variously suggest that th...
In a memory-guided drawing task under blindfolded conditions, we have recently used functional Magnetic Resonance Imaging (fMRI) to demonstrate that the primary visual cortex (V1) may operate as the visuo-spatial buffer, or "sketchpad," for working memory. The results implied, however, a modalit...
│ │ Type DocumentArrayInMemory │ │ Length 16934 │ │ Homogenous...
Cross-modal working memory binding and word recognition skills: how specific is the link?Cross-modal working memory binding and word recognition skills: how specific is the link?Working memorycross-modal bindingepisodic bufferword acquisitionMandarin...
Cross-modal functioning in infancy: IV. A process analysis of infant visual and cross-modal recognition memory: Implications for an amodal code The visual information that specifies three-dimensional objects is often incomplete because objects occlude parts of themselves and other objects. Yet peop.....