From Pixels to Feelings: Aligning MLLMs with Human Cognitive Perception of Images
This work addresses the gap in MLLMs for human-centric AI applications, but it is incremental as it builds on existing models with a new benchmark and training method.
The authors tackled the problem of Multimodal Large Language Models (MLLMs) lacking alignment with human cognitive perception of images, such as memorability or emotional evocativeness, by introducing CogIP-Bench for evaluation and showing that post-training significantly enhances this alignment, with improvements in downstream creative tasks like generating more memorable images.
While Multimodal Large Language Models (MLLMs) are adept at answering what is in an image-identifying objects and describing scenes-they often lack the ability to understand how an image feels to a human observer. This gap is most evident when considering subjective cognitive properties, such as what makes an image memorable, funny, aesthetically pleasing, or emotionally evocative. To systematically address this challenge, we introduce CogIP-Bench, a comprehensive benchmark for evaluating MLLMs on such image cognitive properties. Our evaluation reveals a significant gap: current models are poorly aligned with human perception of these nuanced properties. We then demonstrate that a post-training phase can effectively bridge this gap, significantly enhancing the model's alignment with human judgments. Furthermore, we show that this learned cognitive alignment is not merely predictive but also transferable to downstream creative tasks. By integrating our cognitively-aligned MLLM into an image generation pipeline, we can guide the synthesis process to produce images that better embody desired traits, such as being more memorable or visually appealing. Our work provides a benchmark to measure this human-like perception, a post-training pipeline to enhance it, and a demonstration that this alignment unlocks more human-centric AI.