Coherent Risk Measures
The study links coherent risk measures in finance with uniform integrability in probability theory, using a tool called the folding score of distortion risk measures.
6 shares1 citation todaySource ↗
Quant LetterNo. 44
April 2024, Week 2
90 items across 8 sections, as sent to readers on 10 April 2024. Paper titles open their ML-Quant page; ↗ goes to the source.
- Items
- 90
- Papers
- 57
- Sections
- 8
- Issue
- No. 44
arXiv
Quantitative-finance and ML-for-finance preprints from arXiv.
10 items
Finance6
Social Media Emotions
The research finds that investor emotions expressed on social media can predict daily asset price movements, particularly in low liquidity or high short interest situations.
3 shares4 citations todaySource ↗
The PEAL Method
The paper introduces the PEAL Method, a mathematical framework for structuring securitizations, aimed at improving market transparency, regulatory oversight, and risk management.
3 sharesSource ↗
Improved Multi-Asset Options Bounds
The first article explores the calculation of model-free bounds for multi-asset options, emphasizing the importance of prioritizing relevant information for accuracy and efficiency.
4 sharesSource ↗
Generalized Black-Scholes Equation
The research presents a generalized version of the Black-Scholes model, considering option price dynamics to depend on a measure representing investors' uncertainty.
2 shares2 citations todaySource ↗
Miscellaneous2
Judgment in US Output Growth
The study finds that while US output growth rate predictions are generally unbiased, the use of judgment does not necessarily improve their accuracy.
2 shares1 citation todaySource ↗
Enhanced Electricity Price Forecasting
The study reveals that LQ and elastic net penalty functions provide more precise electricity price predictions than other methods, including the popular LASSO. It also confirms that cross-validation is a useful tool for optimizing parameters.
2 shares13 citations todaySource ↗
Crypto & Blockchain2
Cryptocurrency Volatility
The paper analyzes the factors influencing cryptocurrency volatility from 2020 to 2022, highlighting that positive market returns, positive signed volatility, and negative daily leverage increase price volatility.
8 shares24 citations todaySource ↗
Fee Choice in AMMs
The study explores the workings of arbitrage in decentralized finance automated market makers (AMMs), aiming to understand how AMMs can optimize revenue or minimize losses, and models the dynamics of arbitrage activity.
6 shares1 citation todaySource ↗
SSRN
Working papers in finance and economics from SSRN.
26 items
Quantitative10
Machine Learning for CAT Bond Pricing
The study introduces a machine learning approach for pricing catastrophe bonds, offering more accuracy and robustness than conventional methods, and highlighting key nonlinear relationships between risk factors and bond spreads.
2 sharesSource ↗
Quantum Algorithm for Investment Strategy Testing
A quantum algorithm is shown to outperform traditional algorithms in classifying probability distributions in quantitative finance, offering superior discriminatory power and linear scaling with data samples.
3 shares1 citation todaySource ↗
Generative AI in Asset Management
Hedge funds using generative AI tools like ChatGPT since 2022 have seen increased returns compared to those not using AI.
14 sharesSource ↗
Options Strategies for ETFs
The research examines various option strategies for ETFs, finding their effectiveness varies based on the risk profile of the underlying asset, with some strategies potentially improving risk-adjusted returns.
3 sharesSource ↗
Negative Premium in A-Share Market
The research reveals that stocks with higher volatility have significantly lower returns, an anomaly that can't be explained by market volatility or ambiguity aversion.
2 sharesSource ↗
Reconsidering ML
The article critiques the term 'machine learning', arguing it doesn't accurately represent machine algorithms and suggests a reevaluation of the narratives around these technologies.
18 shares1 citation todaySource ↗
Nonparametric Time Series Bounds
A study explores the properties of empirical risk minimization for time series, focusing on predicting a univariate time series belonging to a class of location-scale parameter-driven processes.
148 sharesSource ↗
Model Risk Management for AI
An article suggests that model risk management can be used in national systemic and cyber risk projects like Project Maven, to transition from AI automation to AI augmentation.
2 sharesSource ↗
Multi-Country Macroeconomic Monitoring
A new model accounting for differences across countries and macroeconomic time series effectively assesses the impact of global shocks on country-level macroeconomic risks.
2 sharesSource ↗
Volatility Risk Pricing
The paper recommends using variance-dependent pricing kernels for option valuation, as they resolve anomalies, fit options well, and provide accurate estimates of equity and variance risk premiums.
2 sharesSource ↗
Financial16
Bank Complexity Measurement Using XAI
A machine learning technique shows a link between the complexity and opacity of banks, with complex firms seeing decreased trading activity.
13 shares1 citation todaySource ↗
Bond and Equity ETF Hidden Liquidity
A study shows hidden liquidity is higher for bond ETFs than equity ETFs, with trading costs and informed trading probability not explaining the difference.
2 sharesSource ↗
Asset Pricing with Cognitive Biases
A deep learning model shows cognitive biases can explain the equity premium puzzle and excess volatility puzzle in asset pricing.
6 sharesSource ↗
Interval Data for FX Volatility Forecasting
The research explores the use of interval-valued data in foreign exchange markets to enhance volatility forecasts, utilizing threshold autoregressive interval models for four major exchange rates.
2 sharesSource ↗
Pairs Trading in the German Stock Market
The study uses various methods to identify and leverage mispricing in the German stock market, revealing that a copula-based method provides a consistent average portfolio return after transaction costs.
2 sharesSource ↗
ESGInvesting: A Psychological Phenomenon
A Psychological Phenomenon: The study expands traditional portfolio selection and asset pricing theory to include ESG investing, introducing two behavioral innovations related to investor preferences and biased judgments about ESG impact and return distributions.
2 shares5 citations todaySource ↗
ESG Impact on Private Equity Fundraising
A study reveals that private equity advisers who disclose negative environmental information in regulatory filings struggle to raise capital.
237 sharesSource ↗
Economic Narratives in Portfolio Management
Machine learning models incorporating economic narratives into market portfolio management have been found to outperform benchmarks, particularly during recessions and high investor sentiment periods.
275 sharesSource ↗
European Institutional Portfolio Comparison
A study of 800 European institutional portfolios shows they are primarily equity risk-focused, have lower carbon emissions but also lower ESG scores compared to the MSCI ACWI equity benchmark.
17 sharesSource ↗
Factor Model for Private Company Valuation
A new method for valuing private companies in private equity, based on actual transactions, can enhance asset allocation and risk management in private markets.
2 sharesSource ↗
High Frequency Risk Pricing for Real Estate Securities
A paper has found that the pricing of default interest rate, liquidity, and excess liquidity risks for securitized commercial real estate securities can yield significant returns, with the first estimates made at intraday frequencies.
2 sharesSource ↗
Customizing Allocation
The article presents a strong framework for customizing asset allocation portfolios, emphasizing the advantages of automation and transparency in portfolio construction.
2 sharesSource ↗
AI Consensus on Pricing
The article discusses the disagreement among machine learning models about which factors affect returns, suggesting that a unified model is impossible with current data.
2 sharesSource ↗
Value of Alternative Data
The article investigates the influence of social media data on mutual fund managers' decisions and performance, indicating that using such data improves future returns and stock-picking skills.
4 sharesSource ↗
Machine Learning for Forecasting
The article shows the superior accuracy of machine learning in predicting dividends, highlighting its effectiveness in complex information structures and potential influence on corporate finance and investment decisions.
2 sharesSource ↗
Subjectivity in Selection
The article delves into the principles of active portfolio selection, emphasizing the potential of subjectivity in improving portfolio performance and the efficiency of a unified active portfolio selection framework.
2 sharesSource ↗
RePEc
Economics working papers from RePEc's NEP field reports.
12 items
Finance6
OR Insights on Finance
The article explores the use of operational research techniques like stochastic programming and machine learning to comprehend risks in financial and economic systems.
18 sharesSource ↗
Sectoral Volatility Contagion
The study examines the structure of risk contagion across sectors, emphasizing the need for accurate identification of this structure for effective regulation.
14 sharesSource ↗
Insider Silence Trading Strategies
The research finds that trading strategies based on insider trading, specifically buying insider purchases and selling insider sales, perform better over extended periods.
12 sharesSource ↗
Portfolio Optimization with Pareto-Dirichlet Method
The article presents a new portfolio optimization method, the Pareto–Dirichlet approach, which surpasses other methods in speed and precision.
11 sharesSource ↗
Forecasting CPI with Multisource Data
The research uses Chinese news and Internet search data to enhance the accuracy and timeliness of CPI forecasting, emphasizing the significance of alternative data in economic downturns or uncertain times.
9 sharesSource ↗
Auto Insurance Risk Management: Value of Vehicles
Value of Vehicles: The article criticizes traditional vehicle insurance pricing methods as outdated, proposing a shift towards flexible price-to-value methods that consider the actual features and values of vehicles.
8 sharesSource ↗
Statistical3
Predicting Systemic Financial Risk
Research suggests that machine learning models and financial stress index can effectively predict systemic financial risk, with stock and money markets being the most influential.
26 sharesSource ↗
Innovative ML Workflow for China's Financial Crisis
A study introduces a multistep workflow using machine learning models and macroeconomic indicators to predict China's systemic financial crisis, accurately identifying six risk periods from 1990 to 2020.
15 sharesSource ↗
Trip Misreporting in Household Travel Survey
The study uses a data-driven method to detect trip misreporting in household travel surveys in Suzhou, China, by incorporating mobile phone signaling data, showing that 23% of total trips were not recorded due to misreporting.
8 sharesSource ↗
Machine Learning3
Data Sensitivity in Machine Learning Reuse
The article explores the difficulties in reusing machine learning applications due to data sensitivity and domain specificity, categorizing applications into four types based on reuse strategies.
23 sharesSource ↗
CostSensitive Machine Learning for Investments
The study employs cost-sensitive machine learning models to predict startup success, potentially reducing investor risk but possibly limiting gains, and proposes ways to improve successful startup detection.
22 sharesSource ↗
US House Price Dynamics
The article presents a new estimator that includes cross-sectional heterogeneity and dependency in machine learning, greatly enhancing the prediction of house prices and detection of housing market bubbles.
8 sharesSource ↗
Machine learning
The general machine-learning papers the letter carried in 2023-25.
9 items
Recently Published4
AutoWebGLM: Better Web Navigation
Better Web Navigation: AutoWebGLM is a new web navigation tool that surpasses GPT-4 in performance, using a unique HTML simplification algorithm and a combined human-AI method to enhance webpage understanding and browser functionality, tested using a bilingual benchmark.
7 shares175 citations todaySource ↗
Evaluating Adversarial Robustness
The study investigates adversarial attacks on Deep Neural Networks for image classification, highlighting the Fast Gradient Sign Method and the Carlini-Wagner approach, and suggests defensive distillation as a defense, effective against FGSM but vulnerable to CW attacks.
7 shares3 citations todaySource ↗
player2vec: Player Behavior in Games
Player Behavior in Games: A new technique for learning hidden user profiles from player behavior data in video and mobile games is presented, utilizing a long-range Transformer model from natural language processing, showing promising results in matching behavior event distribution.
7 shares9 citations todaySource ↗
Spatial Reasoning in LLMs with Visualization
The paper presents Visualization-of-Thought prompting, a method that improves the spatial reasoning abilities of large language models by visualizing their thought processes.
14 shares99 citations todaySource ↗
Historical Trending5
EditFriendly Noise Space
A novel latent noise space has been proposed for denoising diffusion probabilistic models, enabling a variety of image editing operations and perfect image reconstruction.
84 shares308 citations todaySource ↗
NeuralCSA: Causal Sensitivity Analysis
Causal Sensitivity Analysis: The article introduces NeuralCSA, a new neural framework for analyzing causal sensitivity, capable of handling various models, treatments, and queries, and providing accurate causal query bounds.
54 shares20 citations todaySource ↗
Ziya: Data-centric Learning for LLMs
Data-centric Learning for LLMs: The authors present Ziya2, a large language model with 13 billion parameters, focusing on pre-training techniques and data-centric optimization, outperforming other models in multiple benchmarks.
19 shares27 citations todaySource ↗
KV Cache Quantization
KVQuant, a new method for quantizing cached KV activations in large language models, has been developed, allowing the LLaMA-7B model to be served on an 8-GPU system with minimal degradation.
64 shares701 citations todaySource ↗
ChatGLM-Math: Problem-Solving in LLMs
Problem-Solving in LLMs: The authors propose a Self-Critique pipeline to enhance the mathematical problem-solving skills of large language models without affecting their language abilities, showing significant improvements in both areas.
47 shares66 citations todaySource ↗
Papers with code
Papers that shipped their code, from the Papers with Code feed (2023-25).
5 items
Trending5
Unified Structure Learning for Visual Document Understanding
The article highlights the importance of structure information in Visual Document Understanding and presents Unified Structure Learning as a method to improve MLLMs performance.
673 shares
UniTable: Table Structure Recognition Framework
Table Structure Recognition Framework: The factual and quantitative data in tables, along with human-made conventions, make them difficult for machines to interpret.
78 shares
AutoWebGLM: Web Navigation Model
Web Navigation Model: Large language models have difficulty processing real-world webpages due to the diverse actions, extensive HTML text, and complex decision-making involved.
85 shares
Advancing LLM Reasoning with Trees
Eurus is a collection of large language models specifically designed to handle reasoning tasks.
83 shares
Mamba Model Change Detection
The article presents three mechanisms for modeling spatiotemporal relationships in change decoders, which can be integrated with the Mamba architecture for precise change information.
67 shares
GitHub
Repositories the letter featured.
8 items
Finance4
Automated Bug Fixing
SWEagent uses advanced language models like GPT4 to automatically rectify GitHub issues, successfully resolving 12.29% of bugs in the SWEbench test set within 1.5 minutes.
8,511 shares
StockFormer Implementation
The article showcases a PyTorch implementation of the paper StockFormer: Learning Hybrid Trading Machines with Predictive Coding.
62 shares
Machine Learning Models
The article explores quantified machine learning, deep learning models, Alpha factors, quantified resources, and provides related paper codes.
57 shares
Unified Time Series Forecasting
The article presents a unified training method for universal time series forecasting transformers.
283 shares
Trending4
LocalSearch
The article explains LLocalSearch, a local search aggregator that uses LLM Agents for user queries, eliminating the need for OpenAI or Google API keys.
4,226 shares
Python Client
The piece introduces a Python client specifically designed for QuestDB's InfluxDB Line Protocol.
47 shares
Karpathy LLM
The article discusses the process of training LLMs in raw CCUDA, a straightforward programming language.
8,604 shares
Rust Market Simulation
The piece unveils a Rust Market Simulation Library equipped with a Python API.
13 shares
Podcasts
Episodes on markets, quant methods and economics.
3 items
Quantitative3
Alpha Mining with Bogdan
AlphaCube's Patrick Zoro and Bogdan Ivaniuk discuss their alpha mining algorithm, capable of generating 40 million strategies daily on a single CPU.
13 shares
Cliff Asness Investment
Cliff Asness of AQR Capital Management talks about diversification, expectation management, and the use of AI in investment processes in a podcast.
12 shares
Global FX Flow
FX Strategists Patrick Locke, James Nelligan, and Ladislav Jankovic discuss USD & G10 FX trends, the impact of higher oil and weaker CAD data on the dollar, and the potential for CHF depreciation.
7 shares
X / Twitter
Posts from quant researchers on X.
17 items
Quantitative8
Craftsmanship Alpha Revisited
The piece highlights the crucial role of craftsmanship alpha in developing and applying investment styles or beta exposures.
6 shares
Machine Learning and Asset Pricing Notes
The article offers detailed lecture notes on the convergence of machine learning and asset pricing.
5 shares
Quantitative Macro Machine Learning
The article investigates the use of machine learning in the field of quantitative macroeconomics.
5 shares
Jamie Dimon AI augments JPMorgan jobs
JPMorgan has over 2000 AI and machine learning professionals on its payroll.
4 shares
Foundation Models for Time Series Forecasting
The article explores the creation of Chronos, a forecasting model for time series data, built on language model structures.
3 shares
Macro Condition Index Predicts Market Returns
The paper introduces a macro condition index, derived from aggregated macro forecasts, that predicts market trends and behaves countercyclically.
2 shares
Valuation and Security Analysis Lecture Notes
The article presents lecture notes by Bhaskaran Swaminathan on the topics of valuation and security analysis.
2 shares
Weekly Quant Research Recap Subscription
The article introduces a new weekly summary featuring selected quantitative research on investment, macroeconomics, and trading.
2 shares
Miscellaneous9
Bitcoin Halving Forecast
ManGroup suggests the impending Bitcoin halving may influence the cryptocurrency's value.
1 shares
Time Series Forecasting
TimeLLM, a new time series forecasting method that reprograms large language models, has been released on GitHub.
1 shares
Information Asymmetry Modeling
A recent study explores the dynamics between regular-speed traders and high-frequency traders, emphasizing the information imbalance in trading.
1 shares
Transformers for Forecasting
The article provides a guide on using transformers for timeseries forecasting, with a Jupyter notebook example.
0 shares
Matturck 2024 AI Map
Matturck's 2024 AI Market Map showcases 2011 company logos and highlights new companies in different AI fields.
0 shares
New Episode with Asness
The latest episode discusses global diversification, AI, and inefficient markets with Clifford Asness.
0 shares
LLM Papers with Code Collection
The article shares a selection of LLM Papers, often providing accompanying code for reference and implementation.
0 shares
ABFR Webinar Markus K.
Markus K. Brunnermeier from Princeton University is set to speak about Strategic Money and Credit Ledgers at an ABFR seminar, a group interested in AI and big data in economics and finance.
0 shares
Applying Kalman Filter to Financial Data
The author investigates the application of Kalman filters in streamlining price data with minimal lag, but expresses skepticism about the appropriateness of employing a physical motion model in financial data.
5 shares