Persuasion Strategies in Advertisements
Authors: Yaman Kumar, Rajat Jha, Arunim Gupta, Milan Aggarwal, Aditya Garg, Tushar Malyan, Ayush Bhardwaj, Rajiv Ratn Shah, Balaji Krishnamurthy, Changyou Chen
AAAI 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We then formulate the task of persuasion strategy prediction with multi-modal learning, where we design a multi-task attention fusion model that can leverage other ad-understanding tasks to predict persuasion strategies. The dataset also provides image segmentation masks, which labels persuasion strategies in the corresponding ad images on the test split. It can be observed in Table 2 that our model achieves an accuracy of 59.2%, where a correct match is considered if the strategy predicted by the model is present in the set of annotated strategies for a given ad. |
| Researcher Affiliation | Collaboration | 1 IIIT-Delhi, 2 Adobe Media and Data Science Research (MDSR), 3 University at Buffalo |
| Pseudocode | No | The paper provides architectural diagrams and flowcharts (e.g., Figure 5) to illustrate the model, but it does not include any formal pseudocode or algorithm blocks. |
| Open Source Code | Yes | We publicly release our code and dataset at https://midas-research.github.io/persuasion-advertisements/. |
| Open Datasets | Yes | To annotate persuasion strategies on advertisements, we leverage raw images from the Pitts Ads dataset. It contains 64,832 image ads with labels of topics, sentiments, symbolic references (e.g. dove symbolizing peace), and reasoning the ad provides to its viewers (see Appendix:Fig:2 for a few examples). The dataset had ads spanning multiple industries, products, services, and also contained public service announcements. |
| Dataset Splits | Yes | In total, we label 3000 ad-images with their persuasion strategies; and the numbers of samples in train, val and test split are 2500, 250 and 250 resp. |
| Hardware Specification | No | The paper does not provide specific details regarding the hardware used for running the experiments, such as CPU/GPU models, memory, or specific cloud computing instances. |
| Software Dependencies | No | The paper mentions various models and APIs used (e.g., Vision Transformer, BERT, Google Cloud Vision API, SSD, YOLOv5, Dense Cap, BLIP), but it does not provide specific version numbers for software dependencies or libraries required to replicate the experiments (e.g., Python, PyTorch/TensorFlow versions, CUDA versions). |
| Experiment Setup | No | The paper provides details such as image resizing (224x224), patch size (16x16), pre-training dataset (ImageNet 21k), architecture layers (two transformer encoders, hidden dimension 256), choice of activation (sigmoid), and loss function (binary cross-entropy). It also mentions the active learning parameter k=250. However, it does not specify other crucial hyperparameters like learning rate, batch size, number of epochs, or the specific optimizer used for training the model. |