Large Language Models (LLM), which have developed in recent years, enable credit risk assessment through the analysis of financial texts such as analyst reports and corporate disclosures. This paper presents the first systematic review and taxonomy focusing on LLMbased approaches in credit risk estimation. We determined the basic model architectures by selecting 60 relevant papers published between 2020-2025 with the PRISMA research strategy. And we examined the data used for scenarios such as credit default prediction and risk analysis. Since the main focus of the paper is interpretability, we classify concepts such as explainability mechanisms, chain of thought prompts and natural language justifications for LLM-based credit models. The taxonomy organizes the literature under four main headings: model architectures, data types, explainability mechanisms and application areas. Based on this analysis, we highlight the main future trends and research gaps for LLM-based credit scoring systems. This paper aims to be a reference paper for artificial intelligence and financial researchers.

This research presented an empirical investigation of the determinants of the net interest margin in Turkish Banking sector with a particular emphasis on the bank ownership structure. This study employed a unique bank-level dataset covering Turkey`s commercial banking sector for the 2001-2012. Our main results are as follows. Operation diversity, credit risk and operating costs are important determinants of margin in Turkey. More efficient banks exhibit lower margin and also price stability contributes to lower margin. The effect of principal determinants such as credit risk, bank size, market concentration and inflation vary across foreign-owned, state-controlled and private banks. At the same time, the impacts of implicit interest payment, operation diversity and operating cost are homogeneous across all banks

The paper explores the use of Deep Reinforcement Learning (DRL) in stock market trading, focusing on two algorithms: Double Deep Q-Network (DDQN) and Proximal Policy Optimization (PPO) and compares them with Buy and Hold benchmark. It evaluates these algorithms across three currency pairs, the S&P 500 index and Bitcoin, on the daily data in the period of 2019-2023. The results demonstrate DRL's effectiveness in trading and its ability to manage risk by strategically avoiding trades in unfavorable conditions, providing a substantial edge over classical approaches, based on supervised learning in terms of risk-adjusted returns.

We analyze blocks proposed for inclusion in the Ethereum blockchain during 8 minutes on December 3rd, 2024. Our dataset comprises 38 winning blocks, 15,097 proposed blocks, 10,793 unique transactions, and 2,380,014 transaction-block pairings. We find that exclusive transactions--transactions present only in blocks proposed by a single builder--account for 85% of the fees paid by all transactions included in winning blocks. We also find that a surprisingly large number of user transactions are delayed: although proposed during a bidding cycle, they are not included in the corresponding winning block. Many such delayed transactions are exclusive to a losing builder. We also identify two arbitrage bots trading between decentralized (DEX) and centralized exchanges (CEX). By examining their bidding dynamics, we estimate that the implied price at which these bots trade USDC/WETH and USDT/WETH on CEXes is between 3.4 and 4.2 basis points better than the contemporaneous price reported on Binance.

Recognizing the importance of jump risk in option pricing, we propose a neural jump stochastic differential equation model in this paper, which integrates neural networks as parameter estimators in the conventional jump diffusion model. To overcome the problem that the backpropagation algorithm is not compatible with the jump process, we use the Gumbel-Softmax method to make the jump parameter gradient learnable. We examine the proposed model using both simulated data and S&P 500 index options. The findings demonstrate that the incorporation of neural jump components substantially improves the accuracy of pricing compared to existing benchmark models.