---
title: NewsNet-SDF: Stochastic Discount Factor Estimation with Pretrained Language Model News Embeddings via Adversarial Networks
url: https://www.ml-quant.com/papers/arxiv/2505.06864/
site: ML-Quant (https://www.ml-quant.com)
updated: 2026-09-26
license: Summaries CC BY 4.0; links go to the original sources
index: https://www.ml-quant.com/llms.txt
identifier: arXiv:2505.06864
source_url: http://arxiv.org/abs/2505.06864v1
featured: 2025-05-14
citations: 3
topic: LLMs & Text
---


# NewsNet-SDF: Stochastic Discount Factor Estimation with Pretrained Language Model News Embeddings via Adversarial Networks

Discount Factor Estimation: NewsNet-SDF, a deep learning framework, successfully integrates pretrained language model embeddings with financial time series for asset pricing and risk assessment, outperforming traditional models.

- Source: http://arxiv.org/abs/2505.06864v1
- Identifier: arXiv:2505.06864
- Released: 2025-05-11
- First featured: Quant Letter No. 97 (2025-05-14): https://www.ml-quant.com/issues/2025-05-14/
- Citations (Semantic Scholar): 3
- Published in: not yet
- Topic: LLMs & Text

## Related

- [SliceGPT: Compress Large Language Models by Deleting Rows and Columns](https://www.ml-quant.com/papers/arxiv/2401.15024/): Compressing Language Models: The paper introduces SliceGPT, a post-training sparsification scheme for large language models that reduces the network's embedding dimension, maintains high performance, reduces inference computation, and reveals computational invariance in transformer networks.
- [Accurate RNA 3D structure prediction using a language model-based deep learning approach](https://www.ml-quant.com/papers/arxiv/2207.01586/): The paper introduces RhoFold+, a deep learning method that accurately predicts 3D structures of single-chain RNAs from sequences, surpassing existing methods and aiding in RNA structure and function research.
- [Efficient Adversarial Training in LLMs with Continuous Attacks](https://www.ml-quant.com/papers/arxiv/2405.15589/): CAdvUL, a new adversarial training algorithm, enhances the resilience of large language models against adversarial attacks by efficiently calculating attacks in the continuous embedding space.
- [Gecko: Versatile Text Embeddings Distilled from Large Language Models](https://www.ml-quant.com/papers/arxiv/2403.20327/): Compact Text Embeddings: Gecko is a new text embedding model that improves knowledge extraction from large language models, surpassing other models in the Massive Text Embedding Benchmark.
- [SKI Models: Skeleton Induced Vision-Language Embeddings for Understanding Activities of Daily Living](https://www.ml-quant.com/papers/arxiv/2502.03459/): The study presents SKI models, which incorporate 3D skeletons into the vision-language embedding space, using a skeleton-language model to enhance Vision Language Models and Large Vision Language Models.
- [From Deep Learning to LLMs: A survey of AI in Quantitative Investment](https://www.ml-quant.com/papers/arxiv/2503.21422/): The article explores how artificial intelligence, particularly deep learning and large language models, enhances predictive modeling and automation in quantitative investment.