Summary of Scale-invariant Learning-to-rank, by Alessio Petrozziello et al.
Scale-Invariant Learning-to-Rank
by Alessio Petrozziello, Christian Sommeregger, Ye-Sheen Lim
First submitted to arxiv on: 2 Oct 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: None
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The paper presents a learning-to-rank (LTR) model designed for Expedia’s website, which relies on ranking information relevant to users. A key challenge in deploying LTR models is ensuring consistent feature scaling between training and production data. Normalization techniques like standardization and batch normalization can address this issue but are impractical due to latency impacts and distributed real-time inference difficulties. To overcome this, the authors introduce a scale-invariant LTR framework combining deep and wide neural networks to mathematically guarantee scale-invariance at both training and prediction time. The framework is evaluated in simulated scenarios with injected feature scale issues, demonstrating better performance even with inconsistent train-test scaling. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Imagine you’re searching for a hotel on Expedia’s website, but the results are not very good because some important details are hidden or hard to find. This paper tries to fix this problem by creating a new way to rank information that’s more relevant and helpful to users. One big challenge is making sure the model works well in real life, even when the data used to train it is different from the data it will see later. To solve this issue, the authors created a special kind of neural network that can handle different scales of features at both training and prediction time. This means that even if the data is slightly different, the model will still work well. |
Keywords
» Artificial intelligence » Batch normalization » Inference » Neural network
