Publications
2024
Larger Encoders, Smaller Regressors: Exploring Label Dimensionality Reduction and Multimodal Large Language Models as Feature Extractors for Predicting Social Perception
Iván Martín-Fernández, Sergio Esteban-Romero, Jaime Bellver-Soler, Fernando Fernández-Martínez, and Manuel Gil-Martín. 2024. Larger Encoders, Smaller Regressors: Exploring Label Dimensionality Reduction and Multimodal Large Language Models as Feature Extractors for Predicting Social Perception. In Proceedings of the 5th on Multimodal Sentiment Analysis Challenge and Workshop: Social Perception and Humor (MuSe'24). Association for Computing Machinery, New York, NY, USA, 20–27. https://doi.org/10.1145/3689062.3689083
Abstract: Designing reliable automatic models for social perception can contribute to a better understanding of human behavior, enabling more trustworthy experiences in the multimedia on-line communication environment. However, predicting social attributes from video data remains challenging due to the complex interplay of visual, auditory, and linguistic cues. In this paper, we address this challenge by investigating the effectiveness of Multimodal Large Language Models (MM-LLMs) for feature extraction in the MuSe-Perception challenge. Firstly, our analysis of the novel LMU-ELP dataset has revealed high correlations between certain perceptual dimensions, motivating using a single regression model for all 16 social attributes to be predicted for a set of speakers appearing in recorded video clips. We demonstrate that dimensionality reduction through Principal Component Analysis (PCA) can be applied to the label space without a relevant performance loss. Secondly, by employing frozen MM-LLMs as feature extractors, we explore their ability to capture perception-related information. We extract sequence embeddings from the Qwen-VL and Qwen-Audio models and train a Multi-Layer Perceptron over the attention-pooled vectors for each one of the encoders, obtaining a mean Pearson correlation of 0.22 using the average predictions for both models. Our best result of 0.31 is achieved by training the same architecture over the baseline vit-ver and w2v-msp features, which motivates further exploration on how to effectively leverage advanced MM-LLMs as feature extractors. Lastly, a post hoc analysis of our results highlights the limitations of Pearson correlation for evaluating regression performance in this context. In particular, a similar Pearson coefficient can be obtained with two very different prediction sets displaying different levels of variability. We take this result as a call to action in exploring alternative metrics to assess the regression performance for the task.
@inproceedings{10.1145/3689062.3689083,
author = {Mart\'{\i}n-Fern\'{a}ndez, Iv\'{a}n and Esteban-Romero, Sergio and Bellver-Soler, Jaime and Fern\'{a}ndez-Mart\'{\i}nez, Fernando and Gil-Mart\'{\i}n, Manuel},
title = {Larger Encoders, Smaller Regressors: Exploring Label Dimensionality Reduction and Multimodal Large Language Models as Feature Extractors for Predicting Social Perception},
year = {2024},
isbn = {9798400711992},
publisher = {Association for Computing Machinery},
address = {New York, NY, USA},
url = {https://doi.org/10.1145/3689062.3689083},
doi = {10.1145/3689062.3689083},
abstract = {Designing reliable automatic models for social perception can contribute to a better understanding of human behavior, enabling more trustworthy experiences in the multimedia on-line communication environment. However, predicting social attributes from video data remains challenging due to the complex interplay of visual, auditory, and linguistic cues. In this paper, we address this challenge by investigating the effectiveness of Multimodal Large Language Models (MM-LLMs) for feature extraction in the MuSe-Perception challenge. Firstly, our analysis of the novel LMU-ELP dataset has revealed high correlations between certain perceptual dimensions, motivating using a single regression model for all 16 social attributes to be predicted for a set of speakers appearing in recorded video clips. We demonstrate that dimensionality reduction through Principal Component Analysis (PCA) can be applied to the label space without a relevant performance loss. Secondly, by employing frozen MM-LLMs as feature extractors, we explore their ability to capture perception-related information. We extract sequence embeddings from the Qwen-VL and Qwen-Audio models and train a Multi-Layer Perceptron over the attention-pooled vectors for each one of the encoders, obtaining a mean Pearson correlation of 0.22 using the average predictions for both models. Our best result of 0.31 is achieved by training the same architecture over the baseline vit-ver and w2v-msp features, which motivates further exploration on how to effectively leverage advanced MM-LLMs as feature extractors. Lastly, a post hoc analysis of our results highlights the limitations of Pearson correlation for evaluating regression performance in this context. In particular, a similar Pearson coefficient can be obtained with two very different prediction sets displaying different levels of variability. We take this result as a call to action in exploring alternative metrics to assess the regression performance for the task.},
booktitle = {Proceedings of the 5th on Multimodal Sentiment Analysis Challenge and Workshop: Social Perception and Humor},
pages = {20–27},
numpages = {8},
keywords = {affective computing, multimodal large language model., multimodal sentiment analysis, social perception},
location = {Melbourne VIC, Australia},
series = {MuSe'24}
}
2023
Exploring Video Transformers and Automatic Segment Selection for Memorability Prediciton
Martín-Fernández, I., Esteban-Romero, S., Bellver-Soler, J., Gil-Martín, M., & Fernández-Martínez, F. (2023). Exploring Video Transformers and Automatic Segment Selection for Memorability Prediction. Available at https://ceur-ws.org/Vol-3658/paper25.pdf
Abstract: This paper summarises THAU-UPM’s approach and results from the MediaEval 2023 Predicting Video Memorability task. Focused on the generalisation subtask, our work leverages a pre-trained Video Vision Transformer (ViViT), fine-tuned on memorability-related data, to model temporal and spatial relationships in videos. We propose novel, annotator-independent automatic segment selection methods grounded in visual saliency. These methods identify the most relevant video frames prior to conducting memorability score estimation. This selection process is implemented during both training and evaluation phases. Our study demonstrates the effectiveness of fine-tuning the ViViT model compared to a scratchtrained baseline, emphasising the importance of pre-training for predicting memorability. However, the model shows comparable sensitivity to both saliency-based and naive segment selection methods, suggesting that fine-tuning may harness similar benefits from various video segments. These results underscore the robustness of our approach but also signal the need for ongoing research.
@article{martin2023exploring,
title={Exploring Video Transformers and Automatic Segment Selection for Memorability Prediction},
author={Mart{\'\i}n-Fern{\'a}ndez, Iv{\'a}n and Esteban-Romero, Sergio and Bellver-Soler, Jaime and Gil-Mart{\'\i}n, Manuel and Fern{\'a}ndez-Mart{\'\i}nez, Fernando},
year={2023}
}
Video Memorability Prediction From Jointly-learnt Semantic and Visual Features
Iván Martín-Fernández, Ricardo Kleinlein, Cristina Luna-Jiménez, Manuel Gil-Martín, and Fernando Fernández-Martínez. 2023. Video Memorability Prediction From Jointly-learnt Semantic and Visual Features. In Proceedings of the 20th International Conference on Content-based Multimedia Indexing (CBMI '23). Association for Computing Machinery, New York, NY, USA, 178–182. https://doi.org/10.1145/3617233.3617260
Abstract: The memorability of a video is defined as an intrinsic property of its visual features that dictates the fraction of people who recall having watched it on a second viewing within a memory game. Still, unravelling what are the key features to predict memorability remains an obscure matter. This challenge is addressed here by fine-tuning text and image encoders using a cross-modal strategy known as Contrastive Language-Image Pre-training (CLIP). The resulting video-level data representations learned include semantics and topic-descriptive information as observed from both modalities, hence enhancing the predictive power of our algorithms. Our proposal achieves in the text domain a significantly greater Spearman Rank Correlation Coefficient (SRCC) than a default pre-trained text encoder (0.575 ± 0.007 and 0.538 ± 0.007, respectively) over the Memento10K dataset. A similar trend, although less pronounced, can be noticed in the visual domain. We believe these findings signal the potential benefits that cross-modal predictive systems can extract from being fine-tuned to the specific issue of media memorability.
@inproceedings{10.1145/3617233.3617260,
author = {Mart\'{\i}n-Fern\'{a}ndez, Iv\'{a}n and Kleinlein, Ricardo and Luna-Jim\'{e}nez, Cristina and Gil-Mart\'{\i}n, Manuel and Fern\'{a}ndez-Mart\'{\i}nez, Fernando},
title = {Video Memorability Prediction From Jointly-learnt Semantic and Visual Features},
year = {2023},
isbn = {9798400709128},
publisher = {Association for Computing Machinery},
address = {New York, NY, USA},
url = {https://doi.org/10.1145/3617233.3617260},
doi = {10.1145/3617233.3617260},
booktitle = {Proceedings of the 20th International Conference on Content-Based Multimedia Indexing},
pages = {178–182},
numpages = {5},
keywords = {pre-training, media memorability, cross-modal, CLIP},
location = {<conf-loc>, <city>Orleans</city>, <country>France</country>, </conf-loc>},
series = {CBMI '23}
}