{
  "id": 545983,
  "title": "Tips for Increasing Generalizability and Engineering New Imaging-Based Features",
  "url": "/competitions/applications-of-gen-ai-wustl-fall-2024/discussion/545983",
  "author_name": "Robert Jirsaraie",
  "post_date": "2024-11-13T06:16:51.231000",
  "votes": 0,
  "comment_count": 0,
  "views": 0,
  "content": "<p>Hi Everyone,</p>\n<p>I will share a couple general hints, so you all can continue learning and testing your abilities in a productive direction. </p>\n<p>1) Using an ensemble of machine learning models general improves the generalizability of your model. You will find in Jeff's prior competitions the top solutions combined the predictions from multiple predictive algorithem - this approach helped my team win in the fall 2022. Although CatBoost is very practical for this dataset, because it can handle high-dimensional categorical features, you may still want to experiment with alternatives, such as XGBoost. Additionally, you can also derive multiple Catboost algorithem by changing their hyperparameters, such as having one model that is very deep but does not have many trees and another that is very shallow with many trees. There are many hyperparameters and algorithem to test and evaluate.</p>\n<p>2) You can use multimodal models to generate a description of each image of the home and subsequently turn the descriptions into word embeddings that you can add to your model. The effectiveness of this approach will likely come down to how well you can prompt the large language model to describe important elements from the images - that are highly correlate with the home value. I have prompted an example script of how to download and use a multimodal model from huggingfaces. These are powerful models but they require a lot of compute and memory, so running them across all 1200 images might be a challenge. Lastly, focus on engineering features that can help explain the most expensive homes, which are the outliers in this dataset.</p>",
  "messages": [
    {
      "id": 3044179,
      "postDate": "2024-11-13T06:16:51.233Z",
      "content": "<p>Hi Everyone,</p>\n<p>I will share a couple general hints, so you all can continue learning and testing your abilities in a productive direction. </p>\n<p>1) Using an ensemble of machine learning models general improves the generalizability of your model. You will find in Jeff's prior competitions the top solutions combined the predictions from multiple predictive algorithem - this approach helped my team win in the fall 2022. Although CatBoost is very practical for this dataset, because it can handle high-dimensional categorical features, you may still want to experiment with alternatives, such as XGBoost. Additionally, you can also derive multiple Catboost algorithem by changing their hyperparameters, such as having one model that is very deep but does not have many trees and another that is very shallow with many trees. There are many hyperparameters and algorithem to test and evaluate.</p>\n<p>2) You can use multimodal models to generate a description of each image of the home and subsequently turn the descriptions into word embeddings that you can add to your model. The effectiveness of this approach will likely come down to how well you can prompt the large language model to describe important elements from the images - that are highly correlate with the home value. I have prompted an example script of how to download and use a multimodal model from huggingfaces. These are powerful models but they require a lot of compute and memory, so running them across all 1200 images might be a challenge. Lastly, focus on engineering features that can help explain the most expensive homes, which are the outliers in this dataset.</p>",
      "rawMarkdown": "Hi Everyone,\n\nI will share a couple general hints, so you all can continue learning and testing your abilities in a productive direction. \n\n1) Using an ensemble of machine learning models general improves the generalizability of your model. You will find in Jeff's prior competitions the top solutions combined the predictions from multiple predictive algorithem - this approach helped my team win in the fall 2022. Although CatBoost is very practical for this dataset, because it can handle high-dimensional categorical features, you may still want to experiment with alternatives, such as XGBoost. Additionally, you can also derive multiple Catboost algorithem by changing their hyperparameters, such as having one model that is very deep but does not have many trees and another that is very shallow with many trees. There are many hyperparameters and algorithem to test and evaluate.\n\n2) You can use multimodal models to generate a description of each image of the home and subsequently turn the descriptions into word embeddings that you can add to your model. The effectiveness of this approach will likely come down to how well you can prompt the large language model to describe important elements from the images - that are highly correlate with the home value. I have prompted an example script of how to download and use a multimodal model from huggingfaces. These are powerful models but they require a lot of compute and memory, so running them across all 1200 images might be a challenge. Lastly, focus on engineering features that can help explain the most expensive homes, which are the outliers in this dataset."
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "3044179": "Hi Everyone,\n\nI will share a couple general hints, so you all can continue learning and testing your abilities in a productive direction. \n\n1) Using an ensemble of machine learning models general improves the generalizability of your model. You will find in Jeff's prior competitions the top solutions combined the predictions from multiple predictive algorithem - this approach helped my team win in the fall 2022. Although CatBoost is very practical for this dataset, because it can handle high-dimensional categorical features, you may still want to experiment with alternatives, such as XGBoost. Additionally, you can also derive multiple Catboost algorithem by changing their hyperparameters, such as having one model that is very deep but does not have many trees and another that is very shallow with many trees. There are many hyperparameters and algorithem to test and evaluate.\n\n2) You can use multimodal models to generate a description of each image of the home and subsequently turn the descriptions into word embeddings that you can add to your model. The effectiveness of this approach will likely come down to how well you can prompt the large language model to describe important elements from the images - that are highly correlate with the home value. I have prompted an example script of how to download and use a multimodal model from huggingfaces. These are powerful models but they require a lot of compute and memory, so running them across all 1200 images might be a challenge. Lastly, focus on engineering features that can help explain the most expensive homes, which are the outliers in this dataset."
  }
}