RECORD
MindMem: Multimodal for Predicting Advertisement Memorability Using LLMs and Deep Learning
Artificial Intelligence · Submitted 25 Feb 2025
abstract compiled · 2026-09-23 · no language-model review
The authors' main claim Our results highlight the transformative potential of Artificial Intelligence in advertising, offering advertisers a robust tool to drive engagement, enhance competitiveness, and maximize impact in a rapidly evolving market.
The compiled record
The paper's own sentences, arranged. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.
Open the compiled record
The abstract, sentence by sentence
In the competitive landscape of advertising, success hinges on effectively navigating and leveraging complex interactions among consumers, advertisers, and advertisement platforms. These multifaceted interactions compel advertisers to optimize strategies for modeling consumer behavior, enhancing brand recall, and tailoring advertisement content. To address these challenges, we present MindMem, a multimodal predictive model for advertisement memorability. By integrating textual, visual, and auditory data, MindMem achieves state-of-the-art performance, with a Spearman's correlation coefficient of 0.631 on the LAMBDA and 0.731 on the Memento10K dataset, consistently surpassing existing methods. Furthermore, our analysis identified key factors influencing advertisement memorability, such as video pacing, scene complexity, and emotional resonance. Expanding on this, we introduced MindMem-ReAd (MindMem-Driven Re-generated Advertisement), which employs Large Language Model-based simulations to optimize advertisement content and placement, resulting in up to a 74.12% improvement in advertisement memorability. Our results highlight the transformative potential of Artificial Intelligence in advertising, offering advertisers a robust tool to drive engagement, enhance competitiveness, and maximize impact in a rapidly evolving market.
background contribution method result interpretation main claim
| # | sentence | role | the authors' own work | states the main conclusion | strength |
|---|---|---|---|---|---|
| s0 | In the competitive landscape of advertising, success hinges on effectively navigating and leveraging complex interactions among consumers, advertisers, and advertisement platforms. | background 1.00 | 0.04 | 0.02 | asserted 1.00 |
| s1 | These multifaceted interactions compel advertisers to optimize strategies for modeling consumer behavior, enhancing brand recall, and tailoring advertisement content. | background 1.00 | 0.04 | 0.03 | asserted 1.00 |
| s2 | To address these challenges, we present MindMem, a multimodal predictive model for advertisement memorability. | contribution 0.98 | 0.86 | 0.06 | asserted 0.99 |
| s3 | By integrating textual, visual, and auditory data, MindMem achieves state-of-the-art performance, with a Spearman's correlation coefficient of 0.631 on the LAMBDA and 0.731 on the Memento10K dataset, consistently surpassing existing methods. | result 1.00 | 0.96 | 0.30 | asserted 1.00 |
| s4 | Furthermore, our analysis identified key factors influencing advertisement memorability, such as video pacing, scene complexity, and emotional resonance. | result 0.88 | 0.95 | 0.09 | asserted 1.00 |
| s5 | Expanding on this, we introduced MindMem-ReAd (MindMem-Driven Re-generated Advertisement), which employs Large Language Model-based simulations to optimize advertisement content and placement, resulting in up to a 74.12% improvement in advertisement memorability. | contribution 0.77 | 0.95 | 0.17 | asserted 1.00 |
| s6 | Our results highlight the transformative potential of Artificial Intelligence in advertising, offering advertisers a robust tool to drive engagement, enhance competitiveness, and maximize impact in a rapidly evolving market. | interpretation 1.00 | 0.37 | 0.60 | asserted 0.73 |
What was examined, and what was not
- arXiv listing abstract v1
- full text
- figures and tables
- references and citing sentences
- relations to other papers
- P abstract
- S sections
- T body sentences
- C citations
- R cross-paper relations
The paper
Original abstract
In the competitive landscape of advertising, success hinges on effectively navigating and leveraging complex interactions among consumers, advertisers, and advertisement platforms. These multifaceted interactions compel advertisers to optimize strategies for modeling consumer behavior, enhancing brand recall, and tailoring advertisement content. To address these challenges, we present MindMem, a multimodal predictive model for advertisement memorability. By integrating textual, visual, and auditory data, MindMem achieves state-of-the-art performance, with a Spearman's correlation coefficient of 0.631 on the LAMBDA and 0.731 on the Memento10K dataset, consistently surpassing existing methods. Furthermore, our analysis identified key factors influencing advertisement memorability, such as video pacing, scene complexity, and emotional resonance. Expanding on this, we introduced MindMem-ReAd (MindMem-Driven Re-generated Advertisement), which employs Large Language Model-based simulations to optimize advertisement content and placement, resulting in up to a 74.12% improvement in advertisement memorability. Our results highlight the transformative potential of Artificial Intelligence in advertising, offering advertisers a robust tool to drive engagement, enhance competitiveness, and maximize impact in a rapidly evolving market.
Human review
How the literature uses it
What this paper does with the arXiv papers it cites
What it cites · 3 sentences without a confident label
Uses the method or tool of 5 papers
-
For audio embedding, we first extracted audio from videos and fed them into Qwen2 (7B) audio model [ref] , leveraging its last hidden layer to produce audio embeddings.
Qwen2-Audio Technical Report · Multimodal Data Embeddings · 1.00
-
For video embedding, Long Video Assistant (LongVA) model was used to extract visual features from the dataset [ref] .
Long Context Transfer from Language to Vision · Multimodal Data Embeddings · 1.00
-
These descriptions were then processed by the Qwen2 (7B) text model [ref] , which extracted embeddings from the last hidden layer.
Qwen2 Technical Report · Multimodal Data Embeddings · 1.00
Uses data or values from 1 paper
-
To develop and evaluate MindMem, we use two datasets, Long-term Ad MemoraBility DAtaset (LAMBDA) [ref] and Memento10K [ref] , which provide complementary settings for assessing advertisement memorability and general video memorability.
Long-Term Ad Memorability: Understanding & Generating Memorable Ads · Dataset · 0.98
Compares with 3 papers
-
Then, we further compared the performance of MindMem with those of other cutting-edge methods such as Henry [ref] , ViT-Mem [ref] , GPT 3.5 and GPT 4O [ref] .
Image Memorability Prediction with Vision Transformers · Performance of MindMem to Predict Advertisement Memorability · 1.00
-
Then, we further compared the performance of MindMem with those of other cutting-edge methods such as Henry [ref] , ViT-Mem [ref] , GPT 3.5 and GPT 4O [ref] .
GPT-4 Technical Report · Performance of MindMem to Predict Advertisement Memorability · 1.00
-
Then, we further compared the performance of MindMem with those of other cutting-edge methods such as Henry [ref] , ViT-Mem [ref] , GPT 3.5 and GPT 4O [ref] .
Long-Term Ad Memorability: Understanding & Generating Memorable Ads · Performance of MindMem to Predict Advertisement Memorability · 1.00
Discussion
Know something about this paper? Leave a comment, or highlight any passage above to comment on it.