Gaussian Split Tree
The article introduces a new Gaussian Recombining Split Tree (GRST) for valuing securities, improving on traditional binomial trees and aligning with market option prices.
4 sharesSource ↗
Quant LetterNo. 51
158 items across 11 sections, as sent to readers on 28 May 2024. Paper titles open their ML-Quant page; ↗ goes to the source.
Quantitative-finance and ML-for-finance preprints from arXiv.
18 items
The article introduces a new Gaussian Recombining Split Tree (GRST) for valuing securities, improving on traditional binomial trees and aligning with market option prices.
4 sharesSource ↗
The paper showcases Direct Sorted Portfolio Optimization (DSPO), a framework using neural networks to process stock data and construct sorted portfolios, proven effective on various markets.
4 shares1 citation todaySource ↗
The study applies Topological Data Analysis (TDA) to identify extreme stock market events like the 2008 financial crisis and COVID-19 pandemic crashes, highlighting sector-specific impacts.
3 shares18 citations todaySource ↗
The study explores optimal liquidation problems, finding that due to external flows, a player may not fully liquidate their assets.
3 shares1 citation todaySource ↗
The paper introduces a cost-efficient consumption model that uses the Distribution Builder to optimize consumption based on investors' risk preferences and budget constraints.
2 shares9 citations todaySource ↗
The research examines a financial market with risky assets, concluding that transaction costs don't affect equilibrium returns unless there are noise traders.
2 sharesSource ↗
Research indicates that wealth inequality can be reduced and utility increased through a balanced approach to redistribution and consumption morals, promoting a human mutual-aid economy.
3 sharesSource ↗
A study on Kenya's economy from 1991-2015 shows that high-interest external debt negatively impacts GDP growth, suggesting future borrowings should be at lower interest rates.
2 sharesSource ↗
A proposed semi-supervised deep generative framework integrates consumer preferences and external data into product design, allowing companies to create cost-effective, consumer-preferred designs.
2 shares1 citation todaySource ↗
Research on the 2015 Mariana Dam disaster in Brazil shows that environmental changes primarily affected economic outcomes through the production channel.
2 shares1 citation todaySource ↗
A proposed Bayesian decision-theoretic approach optimizes external validity in experiments, revealing efficiency losses when using evidence from randomly-selected sites or those with the largest expected treatment effects.
2 shares1 citation todaySource ↗
The study shows that continuous-time reinforcement learning can be applied to both pure diffusion and jump-diffusion processes, useful in portfolio selection with stock prices modeled as a jump-diffusion.
5 shares2 citations todaySource ↗
The paper introduces a new forecasting method using Generalized Additive Models for precise mid-term hourly electricity load forecasts, improving accuracy and understanding of component influences, beneficial for the power system industry.
3 shares19 citations todaySource ↗
The study presents an Optimal Trading Technique (OTT) for cryptocurrency trading, which uses a dual-objective optimization process to balance profit and risk, achieving a 15.49% annual profit in various market conditions.
5 shares4 citations todaySource ↗
The research suggests four methods to parameterize the DeTEcT framework for modeling wealth distribution in token economies, showing that adding restrictions can create existing wealth distribution models and examining the impact of a dynamic money supply on wealth distribution.
2 shares2 citations todaySource ↗
The study reviews Proof-of-Stake (POS) design options in blockchains, highlighting a balance between improving validator quality and stake quantity for security, and proposes that the best design depends on a platform's specific goals and development stage.
2 shares3 citations todaySource ↗
The study investigates how online trading brokers can increase trading volume by setting a price that encourages buying or selling, using algorithms that consider various feedback situations and valuation distributions.
6 shares1 citation todaySource ↗
The research shows that investors often confuse self-inflated and fundamental returns in active funds, creating a feedback loop that can predict ETF bubbles and crashes, and proposes fund illiquidity as a regulatory measure to prevent self-inflated returns.
4 shares5 citations todaySource ↗
Working papers in finance and economics from SSRN.
40 items
The article proposes a method to separate instantaneous volatility from price process in stochastic volatility models, resulting in a transformed implied volatility skew into a smile.
9 sharesSource ↗
The piece introduces a new neural network-based asset pricing model that includes time-varying volatility dynamics and offers improved predictive accuracy and risk-adjusted returns.
8 sharesSource ↗
The article suggests a new portfolio diversification measure, built from any given risk measure, that meets standard theoretical properties for portfolio diversification.
4 sharesSource ↗
The paper uses a dynamic model to study China's stock index futures market, finding that the futures market's pricing discovery ability is weaker than the spot market's.
4 sharesSource ↗
The article explores the potential of using text data analysis as an alternative measure of interconnectedness between financial institutions, suggesting it can offer valuable insights.
8 sharesSource ↗
A novel regression method for predicting intraday spot volatility outperforms other regression and machine learning techniques in predictive accuracy.
3 shares1 citation todaySource ↗
The accuracy of Machine Learning and Deep Learning in forecasting macroeconomic indicators is compared to the traditional statistical method ARIMA.
2 shares1 citation todaySource ↗
A machine learning model analyzes past loan data to predict the safety of granting loans to individuals, aiming to minimize risk for banks.
3 sharesSource ↗
Decision trees are used to develop intraday trading strategies for individual equities in the NIFTY50 index, potentially improving trading performance and efficiency.
3 sharesSource ↗
The effectiveness of traditional, hybrid, and new subspace classifiers in speaker identification is examined to enhance accuracy rates.
5 sharesSource ↗
The article presents a Bayesian model for measuring GDP growth at high-frequency intervals, which proved effective during the COVID-19 pandemic.
145 sharesSource ↗
The piece suggests that economic downturns directly impact banking industry losses, as evidenced by decreased operational loss recovery rates in large U.S. banks.
126 sharesSource ↗
The article discusses the use of machine-learning algorithms to forecast bank loan risk premium, with SVMs showing the highest accuracy.
15 sharesSource ↗
The study uses a panel dataset to predict poverty status in Nigeria, finding that demographic and housing indicators can accurately predict poverty in 80% of cases.
99 sharesSource ↗
The paper presents a method for identifying causal interactions between variables, which has been validated in predicting stock return and volatility in financial markets.
3 sharesSource ↗
The article corrects a misconception in Kalman's estimator by discussing two types of orthogonality applied to data and time samples.
2 sharesSource ↗
The paper proposes a new method to monitor systemic risks from nonbank financial institutions by integrating multiple data sources.
2 sharesSource ↗
The article introduces a more stable and memory-efficient machine learning method, the Batchstochastic Subgradient method, tested using structured query language.
2 sharesSource ↗
The article outlines new global rules on the patentability of artificial intelligence inventions issued by major patent offices.
4 shares1 citation todaySource ↗
The study explores how the investor panel feature on equity crowdfunding platforms can enhance the fundraising performance of startups.
2 sharesSource ↗
The study proposes a unique trading strategy using the Volatility Risk Premium (VRP) for ETFs, resulting in an average annual return of 20.79%.
5 sharesSource ↗
The article highlights the increasing use of net asset value debt (NAV Debt) in private equity buyouts, shifting liabilities to the fund level.
8 sharesSource ↗
The research examines how research design choices can significantly influence the profitability of Machine learning investment strategies.
6 sharesSource ↗
The paper reveals that listed real estate (LRE) can effectively hedge against inflation in the long term, regardless of economic conditions.
4 shares1 citation todaySource ↗
The article suggests that pension fund investment teams can enhance retirement offerings by prioritizing member outcomes.
7 sharesSource ↗
The finance industry is using business analytics for predictive modeling, fraud detection, customer personalization, and operational optimization, thanks to technological and data advancements.
2 shares1 citation todaySource ↗
The Capital Market Assumptions document provides expected returns, volatility, and correlation estimates for various fixed income assets, taking into account inflation and foreign exchange rate changes.
3 sharesSource ↗
A study of 26 sustainable finance taxonomy frameworks found that few support the transition to carbon neutrality, with many lacking a dynamic approach or only targeting specific financial products.
5 shares6 citations todaySource ↗
SquaredLabs is transforming derivatives trading by using power perpetuals to remove liquidation risks, providing traders of all levels with advanced financial instruments for better risk management and capital efficiency.
2 sharesSource ↗
The article proposes a portfolio management strategy that uses the VIX volatility index to determine leverage, leading to more stable weights, less rebalancing, and higher returns considering transaction costs.
935 sharesSource ↗
The piece presents a method for simulating asset price and variance under the Hull and White stochastic volatility model, useful for generating unbiased estimates for derivatives instruments pricing.
45 sharesSource ↗
The research explores the occurrence, causes, and effects of cheap stock - equity-based compensation granted pre-IPO at a lower share price, and its influence on firms' post-IPO behavior.
152 sharesSource ↗
The article examines the effect of Zero-Day-to-Expiration (0DTE) options trading on stock market volatility, revealing that increased 0DTE options trading significantly boosts volatility.
2 sharesSource ↗
The article proposes a model incorporating mean reversion, stochastic volatility, convenience yield, and jump clustering features of commodity markets, offering a method to price geometric and arithmetic Asian options.
52 sharesSource ↗
The article presents PINstimation, an R package for estimating the probability of informed trading models, using data from 58 Swedish stocks as examples.
349 sharesSource ↗
The study examines climate risks in the European stock market, suggesting these risks are factored into prices and proposing a green rating system for non-disclosing companies.
261 sharesSource ↗
The research explores how credit market sentiment affects the real economy through the term premium, explaining its role in predicting recessions and influencing future output growth.
13 sharesSource ↗
The article investigates the impact of USDA announcements on commodity options, noting significant trading volume changes around the release of monthly agricultural reports and evidence of informed trading.
3 sharesSource ↗
The article introduces a geometric method for incorporating investor views in portfolio construction, offering more flexibility than traditional Black-Litterman model-based approaches.
10 sharesSource ↗
Economics working papers from RePEc's NEP field reports.
21 items
The study reveals gold as a more effective hedge than oil for Pakistani stocks, particularly after COVID-19, and advises investors to diversify their portfolios.
16 sharesSource ↗
The research shows that option signals from leveraged ETFs can predict the performance of the underlying ETFs, especially during economic downturns, and can be utilized to create a profitable trading strategy.
15 sharesSource ↗
The study finds ensemble boosting tree models more effective than traditional ones in predicting volatility in China's crude oil futures, with different variables contributing differently in various models.
20 sharesSource ↗
The research shows that the Large Language Model, ChatGPT, selects more diverse and high-performing assets in portfolio management than random selection.
18 sharesSource ↗
The study uses machine learning to predict employee turnover, identifying key factors like education and payment tier, and finds the Decision Tree model most accurate.
17 sharesSource ↗
The research uses machine and deep learning models to forecast metal futures, finding their efficiency varies based on metal choice, sample period, and inputs.
17 sharesSource ↗
The study finds logistic regression models more effective than random forest models in predicting rainfall-induced landslides in India, especially when using antecedent precipitation data.
14 sharesSource ↗
The FS-RS-ML framework, using machine learning to predict credit risk in supply chain finance for small and medium-sized enterprises, has proven superior in tests using Chinese data.
25 sharesSource ↗
Machine learning can enhance trade direction classification in corporate bond markets, with trade timing and information environment impacting the accuracy of existing rules.
22 sharesSource ↗
A machine learning framework has been proposed to identify delayed recognition research, with a study in chemistry confirming its effectiveness and adaptability.
21 sharesSource ↗
Age-period-cohort models offer a method for comprehensive credit risk modeling across a company, optimizing underwriting based on economic scenarios.
20 sharesSource ↗
A machine learning regression model using remote sensing data can estimate economic performance in data-poor areas, as demonstrated in a South Sudan study.
19 sharesSource ↗
Machine learning can predict global stock returns, with accuracy decreasing over time, but using firm-specific data can improve long-term predictions.
18 sharesSource ↗
A new approach suggests that the lending platform Prosper likely uses an ensemble algorithm to adapt to changes in borrower and lender behavior.
16 sharesSource ↗
KNN vs Naive Bayes: The K-Nearest Neighbours algorithm is more accurate than the Naive Bayes method in predicting storm warnings, with an accuracy rate of 68.20%.
12 sharesSource ↗
Predictive models can forecast the future of life insurance premium payments, identifying policyholders less likely to pay and the impact of surrender fees.
12 sharesSource ↗
The German Federal Statistical Office is studying if machine learning can predict pension taxation data to speed up the publication of these statistics.
12 sharesSource ↗
A new deep learning strategy for financial hedging has been created, offering improved risk management and an average annual economic benefit of 1.21 million CNY for a typical Chinese aluminum firm.
16 sharesSource ↗
A deep learning framework for predicting Bitcoin market crashes has been proposed, surpassing six other machine learning models by 15.8% in balanced accuracy.
14 sharesSource ↗
Evidence from Pakistan Stock Exchange: Research on the Pakistan Stock Exchange–Karachi Meezan Index (PSX–KMI) All Share Index shows that multifactor models can accurately price Shariah-compliant sustainable equity portfolios, with a few exceptions.
16 sharesSource ↗
A study on European equity prices reveals that carbon risk is a systematic risk factor, with firms that are less carbon-intensive providing higher returns, indicating that investors can control carbon risk through stock returns data.
14 sharesSource ↗
The general machine-learning papers the letter carried in 2023-25.
19 items
The article introduces GCD, a new technology that can create videos from any perspective without needing depth or 3D scene geometry, with potential applications in robotics and driving environments.
79 shares130 citations todaySource ↗
The paper shows that Differentiable Annealed Importance Sampling (DAIS) provides more accurate uncertainty estimates than traditional variational inference by minimizing the symmetrized KL divergence.
15 shares3 citations todaySource ↗
The authors demonstrate that score-based generative models are resilient to errors, using the Wasserstein uncertainty propagation theorem to explain how learning errors affect the model's quality.
11 shares15 citations todaySource ↗
The article presents a new ray tracing-based method to enhance Neural Radiance Fields' (NeRFs) rendering of highly reflective objects, showing superior performance and photorealistic results in real-world scenes.
10 shares41 citations todaySource ↗
PuzzleAvatar is a new model that creates 3D avatars from personal photos, eliminating the need for body and camera pose estimation, and performs better than existing models in terms of accuracy and robustness.
10 shares29 citations todaySource ↗
The LD3 algorithm, which can identify evidence of direct discrimination in a polynomial time, offers a more efficient method for causal fairness analysis in complex decision systems.
6 shares8 citations todaySource ↗
Semantica, an image-conditioned diffusion model, can generate new images based on the semantics of a conditioning image, without the need for fine-tuning.
6 sharesSource ↗
The study uses sparse autoencoders to explore the multi-dimensional nature of language model representations in GPT-2 and Mistral 7B, and identifies tasks where these features solve computational problems.
6 shares207 citations todaySource ↗
The paper reveals artifacts in Vision Mamba's feature maps and introduces Mamba-R, a new architecture with register tokens that improves feature maps, performance, and scalability.
5 shares58 citations todaySource ↗
The article introduces MNIST-1D, a cost-effective alternative to traditional deep learning benchmarks, ideal for quick prototyping and low-budget research.
614 shares35 citations todaySource ↗
The article suggests a self-play-based method, SPPO, for language model alignment, which can effectively enhance the likelihood of the selected response and reduce that of the discarded one.
309 shares278 citations todaySource ↗
The article introduces Cross-Layer Attention (CLA), a new attention design that minimizes the key-value cache size, allowing for longer sequence lengths and larger batch sizes during inference.
206 shares140 citations todaySource ↗
The article presents BiomedParse, a biomedical base model for image parsing, capable of performing segmentation, detection, and recognition for various object types across multiple imaging modalities, enhancing accuracy and efficiency in biomedical image analysis.
137 shares193 citations todaySource ↗
OmniGlue, a new image matcher that performs better on unseen image domains than previous models, is introduced in this paper.
67 shares109 citations todaySource ↗
This study introduces a unified approach to improving the efficiency of convolutional neural networks, resulting in more accurate and less costly models.
35 shares5 citations todaySource ↗
The research presents an algorithm, energy rank alignment (ERA), that optimizes the generation of molecules with desired properties, demonstrating strong performance across diverse chemical spaces.
25 shares7 citations todaySource ↗
The study proposes a new approach using deep learning to ensure fair auctions, providing strong statistical guarantees and preventing manipulation of results.
19 sharesSource ↗
The paper suggests a method to improve offline reinforcement learning by limiting out-of-distribution actions during training, showing improved performance on D4RL benchmarks.
14 shares6 citations todaySource ↗
Papers that shipped their code, from the Papers with Code feed (2023-25).
8 items
Webcrawling is an essential research tool for both computational and non-computational scientists.
25,232 shares
Training large language models using reinforcement learning from human feedback presents coordination challenges across four models.
1,338 shares
LightAutoML, an AutoML system, is designed to meet the specific needs of a large European financial services company.
945 shares
DIAMOND, a reinforcement learning agent, is developed in response to a significant change in the prevailing paradigm.
96 shares
The article introduces a novel method that computes and saves the KVs of certain layers, which decreases memory consumption and boosts inference speed.
68 shares
The piece discusses the effectiveness of lowrank adaptation in fine-tuning large language models.
67 shares
The article highlights how regularization enhances performance on benchmarks such as Tiny Stories and SuperGLUE, while also reducing the models' linearity.
38 shares
The article reviews the remarkable performance of Multimodal Large Language Models (MLLMs) in tasks like visual question answering and visual understanding over the past year.
30 shares
Repositories the letter featured.
10 items
AlgoBulls offers a collection of official algorithmic trading strategies.
121 shares
Advanced educational resources for statistics and machine learning studies are available.
70 shares
A collection of resources on using deep learning for time series forecasting is compiled.
2,460 shares
The article discusses the first set of experiments performed on the KolmogorovArnold Network, a system designed for reinforcement learning.
194 shares
TranspileAIivy: This article provides a detailed review of the Unified AI Framework, explaining its features and applications.
14,028 shares
The article provides a detailed, step-by-step guide on how to implement the llama3 algorithm using matrix multiplication.
5,772 shares
KxSystemskdbaisamples: This piece offers practical examples and guidance for developers who are using the KDB database system.
54 shares
StockFundXCrawler: The article describes a simple, efficient project for tracking and analyzing A-shares in the Chinese stock market.
258 shares
Industry news: funds, hiring, markets and regulation.
19 items
Deutsche Bank hires a seasoned algorithmic trading technologist.
6 shares
According to the UBS Global Family Office Report 2024, one-third of global family offices use hedge funds for diversification.
6 shares
Cliff Asness, co-founder of AQR Capital Management, sees the emergence of generative AI as an evolution, equating it to advanced statistics.
5 shares
Joanna Alpert is promoted by Bridgewater Associates to head its new Total Portfolio Strategies unit, responsible for creating and managing new investment products.
4 shares
Bloomberg reports a surge in investment in dispersion, a strategy used by bank trading desks, in the post-pandemic era.
4 shares
Global hedge funds are increasing their Chinese equities holdings for the fourth week in a row, expecting a market rebound, according to Goldman Sachs’ prime brokerage team.
3 shares
New York lawmakers may alter a law that imposes a 9% interest rate on defaulted sovereign emerging-market bonds to avoid disputes like the Elliott Investment Management-Argentina case.
3 shares
Sanjay Shah, founder of Solo Capital, denies accusations of a £1.44bn dividend tax fraud case against Denmark, claiming he relied on others for tax and legal advice.
3 shares
Jane Street Citadel & offers the highest paying internships in the finance sector.
3 shares
Goldman Sachs reports hedge funds are decreasing their investments in large-cap tech stocks, instead focusing on firms benefiting from AI technology.
2 shares
Michael Graveline is appointed as Senior Vice President at Aristotle Capital Management, with a role to build and support partnerships with RIAs and bank trust advisors.
2 shares
Gryphon Fund Group chooses FundGuard as its main accounting book of record for its open-end fund client base.
2 shares
Sydney-based hedge fund Caledonia continues to heavily invest in Zillow Group, anticipating a recovery following a 26% share decrease this year.
2 shares
Two funds formerly managed by Odey Asset Management were among the top 10 best performing funds globally in April, as per Société Générale data.
2 shares
The Sohn Hong Kong Investment Leaders Conference spotlighted stocks from Japan, South Korea, and the Rio Tinto Group, a leading metals and mining company.
1 shares
The Abu Dhabi Investment Authority has hired another ex-banker to its team.
1 shares
Hedge fund manager Pierre Andurand anticipates a significant increase in copper prices due to growing demand and dwindling global reserves.
1 shares
Renowned Australian stock investor Robert Luciano has offloaded large shares in two ASX-listed funds he established, ahead of launching his new venture.
1 shares
The article explores the benefits of participating and striving to win in Citadel's Datathon competition.
0 shares
Episodes on markets, quant methods and economics.
10 items
Michael Melissinos shares his experience in trend following trading and emphasizes the need to eliminate biases for successful trades in a podcast.
15 shares
Stephen Dulake, Andrew Crook, and Samantha Azzarello share insights from the J.P. Morgan Global Markets Conference, focusing on the implications for credit markets.
9 shares
Alex Blostein from Goldman Sachs Research discusses the challenges and opportunities in the private equity industry, with input from Mike Nickols and Gina Lytle on current deal trends.
9 shares
Jonny Goulden and Saad Siddiqui explore the effects of recent market changes on the EM fixed income asset class in a podcast episode.
8 shares
Ben Bennett discusses the effects of stricter lending conditions on the US consumer, the disparity in equity and credit markets, and the increasing cost of red metal.
8 shares
In a podcast, Mustafa from Macro Hive explains misconceptions about current rates, the increase in household wealth, and the importance of levels over changes.
7 shares
Greg Neufeld talks about his interest in 'data flywheel' companies and the transition into venture capital in a podcast interview.
6 shares
Samar Sen discusses the development of institutional tools for digital assets and the tools being developed by Talos in a podcast.
4 shares
Rob Almeida and Genevieve Gilroy discuss the changing dynamics in the consumer staples sector and the significance of global research collaboration in a podcast.
4 shares
Shikha Chaturvedi discusses the factors behind recent price surges in US and European gas markets, including supply reduction and the Ukraine-Russia transit deal, in a podcast.
3 shares
Posts from quant and economics blogs and newsletters.
6 items
The Negative Volume Index (NVI) is a unique tool used in technical trading analysis.
5 shares
A clear trading strategy is essential for successful trading in the S&P 500 Index (SPX).
4 shares
Some traders use moon cycle phases to time the market and make profitable trades.
2 shares
The article examines the application of the Rainbow Oscillator in financial market analysis.
1 shares
The article delves into different technical indicators, including recent ones, used in trading.
1 shares
The article represents the 47th weekly installment of a particular series.
1 shares
Talks, lectures and tutorials.
2 items
The video outlines the three key sectors of quantitative finance - trading, portfolio management, and risk management - and the necessary skills for each role for computer science graduates.
0 shares
The Journal of Finance held a webinar showcasing the presentation and discussion of the 2023 First Prize Brattle Group Prize-winning research paper on the topic of Specialization in Bank Lending.
13 shares
Posts from quant researchers on X.
5 items
The author discusses their appreciation for a trendfollowing episode involving Quantica Capital and Choffstein, focusing on a specific paper.
1 shares
Despite a strong start for hedge funds in 2024, the author explores potential risks as indicated by ManGroup.
1 shares
Owen Lamont analyzes the effects of increasing borrowing costs on the stock loan market.
0 shares
The article highlights the long-term effectiveness of trend-following strategies, particularly when equities are underperforming.
0 shares
The article features an enlightening interview with financial expert, Clifford Asness.
0 shares