Reviews: Modulating early visual processing by language

Overall Impression: I think this paper introduces a novel and interesting idea that is likely to spark future experimentation towards multi-modal early-fusion methods. However, the presentation and the writing could use additional attention. The experiments demonstrate the effectiveness of the approach on multiple tasks though they are a bit narrow to justify the proposed method outside of the application domain of vision language. I think further iterations on the text and additional experiments with other model architectures or different types of multi-modal data would strengthen this submission. Strengths: I like the neurological motivations for the CBN approach and appreciate its simplicity.