SAM: Semantic Attribute Modulation for Language Modeling and Style Variation
This work addresses the need for more controllable and interpretable text generation in natural language processing, though it appears incremental as it builds on existing attribute-based methods.
The paper tackles the problem of generating interpretable texts conditioned on document attributes by proposing Semantic Attribute Modulation (SAM), which embeds attributes like titles and categories into a hidden semantic space and uses an attribute attention mechanism, resulting in improved text generation on several datasets and demonstrating style variation for lyric generation.
This paper presents a Semantic Attribute Modulation (SAM) for language modeling and style variation. The semantic attribute modulation includes various document attributes, such as titles, authors, and document categories. We consider two types of attributes, (title attributes and category attributes), and a flexible attribute selection scheme by automatically scoring them via an attribute attention mechanism. The semantic attributes are embedded into the hidden semantic space as the generation inputs. With the attributes properly harnessed, our proposed SAM can generate interpretable texts with regard to the input attributes. Qualitative analysis, including word semantic analysis and attention values, shows the interpretability of SAM. On several typical text datasets, we empirically demonstrate the superiority of the Semantic Attribute Modulated language model with different combinations of document attributes. Moreover, we present a style variation for the lyric generation using SAM, which shows a strong connection between the style variation and the semantic attributes.