Quant LetterNo. 56
156 items across 10 sections, as sent to readers on 10 July 2024. Paper titles open their ML-Quant page; ↗ goes to the source.
Quantitative-finance and ML-for-finance preprints from arXiv.
21 items
The study improves the pricing accuracy of Asian options by deriving the subleading correction to the implied volatility in the Black-Scholes model, which is determined by the large deviations property for the time-average of the geometric Brownian motion.
6 shares3 citations todaySource ↗
A model is proposed that maximizes daily trading profit and minimizes end-of-day inventory penalization, using a partially observable stochastic control problem to manage unobserved toxicity in client orders.
6 shares7 citations todaySource ↗
The COGARCH and Barndorff-Nielsen and Shephard models are adapted to a Markov-switching environment, allowing for sudden volatility jumps at regime switch times and capturing key characteristics of financial time-series.
5 shares1 citation todaySource ↗
The second-order Esscher pricing concept is introduced for continuous-time models, characterizing the second-order Esscher densities using pointwise equations and relating them to the Delbaen-Haenzendonck's risk-neutral measure.
5 shares1 citation todaySource ↗
Conditional Autoregressive Expected Shortfall: The Conditional Autoregressive Expected Shortfall (CAESar) methodology is introduced for estimating Value at Risk (VaR) and Expected Shortfall (ES), providing a more comprehensive measure of tail risk and outperforming existing regression methods in forecasting performance.
4 shares1 citation todaySource ↗
The article presents a method that merges expert models using diffusion processes and deep learning, specifically for combining implied volatility smiles models from various datasets.
4 shares4 citations todaySource ↗
The paper investigates a quantum method for tracking the progression of traded market prices with uncertain volatility, utilizing projective measurements and Monte-Carlo simulations.
3 sharesSource ↗
The study presents the Dynamic Modularity-Spectral Algorithm (DynMSA) for identifying stock clusters, uncovering hidden market structures and enhancing portfolio allocation, specifically applied to the S&P 500 constituents.
3 sharesSource ↗
The research proposes a method for estimating high-dimensional covariance matrices in latent factor models by clustering residual series, focusing on the idiosyncratic component.
3 shares3 citations todaySource ↗
Stock Prediction: The paper introduces a graph neural network-based convolutional neural network model for predicting stock market prices, using custom feature engineering on diverse data sources.
3 shares11 citations todaySource ↗
AI algorithms using Q-learning can encourage silent collusion in two-sided markets, increasing profits; a penalty term in the algorithm could reduce this.
7 shares4 citations todaySource ↗
Machine learning models are more accurate than traditional methods or expert forecasts in predicting China's quarterly GDP growth, especially during stable economic periods.
5 shares9 citations todaySource ↗
The VF-PGI-Spectral algorithm effectively solves continuous action, infinite-horizon dynamic models in both single and multi-agent games.
3 sharesSource ↗
Local increases in housing prices impact renters' mobility and spending differently; those who move tend to increase consumption and buy homes and cars, while those who stay are largely unaffected.
2 sharesSource ↗
The article introduces a new theory for principal component analysis (PCA) under the weak factor model. This theory accounts for cross-sectional dependent components and provides finite-sample characterizations for estimation error and statistical inference uncertainty level, improving upon previous research.
2 shares7 citations todaySource ↗
The article discusses strategies for managing the risk of impermanent loss in Decentralised Exchanges. It introduces a concept called IL protection claim and provides methods for valuing and managing this claim using arbitrage-based techniques and valuation formulas.
7 shares14 citations todaySource ↗
The article explores the optimization of hedging with variational preferences under convex risk measures, discussing dual representation, optimality, and indifference pricing conditions.
6 sharesSource ↗
The study examines the geographic imbalance of deposits and loans, proposing a method to evaluate the impact of branch networks, market power, and scope economies on this imbalance.
6 shares15 citations todaySource ↗
The paper presents the Minimum Price Markov Game as a model for studying market fairness and regulation in public auctions, emphasizing the challenge of achieving algorithmic tacity coordination.
6 shares2 citations todaySource ↗
The article uses a balance sheet-based model to analyze run risk in banking systems, using the Silicon Valley Bank meltdown as a case study to show how changes in funding and asset composition can increase vulnerability.
5 shares2 citations todaySource ↗
The study suggests an efficient method for pricing derivatives tied to asset baskets, using a copula model to capture the dependency structure among assets and a local volatility model for pricing.
5 sharesSource ↗
Working papers in finance and economics from SSRN.
25 items
The study introduces an equity investment strategy that uses artificial intelligence, multi-factor models, and financial indicators to predict returns and mitigate risks.
4 sharesSource ↗
The paper presents a deep delta hedging framework for options, using neural networks to improve hedging performance by learning the residuals between the hedging function and the implied Black-Scholes delta.
3 sharesSource ↗
The research suggests an adaptive hybrid model using automatic machine learning for short-term PET prediction, showing that the effectiveness of neural networks varies depending on the data sources used.
2 sharesSource ↗
The study applies a validated simulation of the plastic injection process and machine learning to estimate the necessary clamping force, optimizing key parameters and reducing defects and energy use in the pipes and fittings industry.
2 sharesSource ↗
The paper introduces a method for optimizing the topology of three-dimensional frames under static seismic loads using a hierarchical graph-based machine learning model, proving its effectiveness over traditional methods in optimizing large three-dimensional building frames.
2 sharesSource ↗
The study uses both experimental and machine learning methods to examine how the structure of zeolites affects the isomerization of 1octene.
2 sharesSource ↗
The paper suggests using the Alpha Plane concept and a new algorithm to improve transformer protection in electric power systems, with simulations conducted on MATLABSIMULINK.
2 sharesSource ↗
The author integrates financial frictions from company default choices into an open-economy model to create realistic exchange rate patterns, solving the BackusSmith puzzle.
2 sharesSource ↗
The research introduces a new artificial intelligence-machine learning pipeline for automated target recognition using topological data analysis from a sensing grid’s multimodal data.
3 sharesSource ↗
The paper presents a multiagent equilibrium model with discrete cash flow in an incomplete market, deriving the equilibrium interest rate and market price of risk for government bonds.
3 sharesSource ↗
The article explores methods for hedging against the loss of liquidity in Decentralised Exchanges, using Uniswap V2 and V3 protocols, and proposes a new risk management claim.
3 sharesSource ↗
The study examines the use and acceptance of three epistemic adverbs in Colombian Spanish, finding that their acceptability varies based on the type of hedging and the adverb used.
3 sharesSource ↗
The piece warns about the risks of variable leveraged tokens on Binance crypto exchanges, due to the lack of information on potential losses and the absence of margin calls.
3 sharesSource ↗
The author presents an asset pricing model that shows long-term investors hedge against reinvestment risk, leading to significant positive returns for portfolios exposed to this risk.
2 sharesSource ↗
The article discusses the importance of accurate machine learning models for predicting weather-induced power outages, to help utility companies minimize damage to the power system.
2 sharesSource ↗
Research indicates that gold, treasury bills, and interbank rates can act as risk-free assets in various countries, but no such asset exists universally for the UK, US, China, Japan, and India.
2 sharesSource ↗
Private equity investment boosts digital technology investments in portfolio firms, especially when investors are digitally savvy, narrowing the digital investment gap between public and private firms.
2 sharesSource ↗
In emerging economies with limited foreign exchange markets, a mix of monetary policy and foreign exchange intervention can stabilize the economy, as demonstrated in Korea.
2 sharesSource ↗
Strategic HR practices like performance management and transparent recruitment significantly enhance good governance principles such as accountability, transparency, and efficiency in Zamboanga City, Philippines.
2 sharesSource ↗
A study shows that return on equity greatly influences the Sustainable Growth Rate (SGR) and stock prices, suggesting that profitability and effective equity management are crucial for sustainable growth.
2 sharesSource ↗
A study indicates that user behavior changes on the Reddit forum Wallstreetbets can predict market volatility, with less influential users having a greater impact.
13 sharesSource ↗
Research indicates that companies with a higher perceived cost of capital earn more returns and invest less, leading to a misallocation of capital that reduces productivity.
3 sharesSource ↗
A study recommends removing assets before calculating portfolio weights to best mitigate estimation risk in portfolio selection, outperforming other sparse methods.
5 sharesSource ↗
Offshore-domiciled European mutual funds in Ireland or Luxembourg outperform onshore ones due to more active management and a more sophisticated investor base.
3 sharesSource ↗
The study shows that when fund managers are compared more frequently to industry standards, it leads to an increase in a stock's risk level (CAPM β). This causes companies to decrease investment, hold more cash, and give more payouts to shareholders. These effects last for at least seven years.
2 sharesSource ↗
Economics working papers from RePEc's NEP field reports.
30 items
The research examines the fluctuation in commodity returns and financial market index, suggesting that different models work best for different assets, improving risk management strategies.
31 sharesSource ↗
The article investigates the cryptocurrency market, revealing that high-frequency positive returns increase price volatility, unlike traditional assets like stocks.
25 sharesSource ↗
The study explores the volatility of the Asian stock market in relation to Bitcoin and global crude oil prices, showing volatility clustering and varying spillover effects.
24 sharesSource ↗
The study looks at the impact of market conditions, volatility, and liquidity shocks on arbitrage profits during pre-COVID and COVID periods, indicating that high volatility and low liquidity make arbitrage unfeasible.
23 sharesSource ↗
The research proposes a theory for investors to use alternative data such as social media and pandemic information to predict stock trends and guide investment decisions.
22 sharesSource ↗
The study recommends using an eight-factor model in corporate valuation to reduce estimation errors and increase precision over traditional methods.
21 sharesSource ↗
The article presents a method to calculate the likelihood of reaching a specific profitability goal with strategic portfolios requiring regular rebalancing, based on data from 2004 to 2021.
16 sharesSource ↗
The research reveals that Bitcoin return volatility negatively affects the returns of Shariah-compliant stocks, suggesting implications for diversification strategies of investors.
12 sharesSource ↗
The study uses machine learning to improve the estimation of Loss Given Default (LGD) in situations with limited cash-flow data, using a European mortgage portfolio.
25 sharesSource ↗
The research finds that using machine learning to select a subset of stocks can yield higher profits than the S&P 500, despite increased risk.
24 sharesSource ↗
The study suggests a method combining decision trees and time series modeling to predict play calls in the National Football League based on various factors.
23 sharesSource ↗
The paper introduces a model for optimizing loans across four retail asset classes, considering regulatory and capital constraints, and compares the optimized portfolio to the original for potential benefits.
20 sharesSource ↗
The article proposes a new method in machine learning that combines the conditional predictive impact framework with sequential knockoff sampling to better distinguish between marginal and conditional measures in feature importance.
15 sharesSource ↗
The paper introduces a machine learning approach to predict the recovery rate of non-performing loans, aiming to minimize the lemon discount by accurately pricing the risk component of information asymmetry between banks and investors.
14 sharesSource ↗
The article presents a novel machine learning framework for long-term mortality forecasting, improving prediction accuracy and addressing the issue of diminishing patterns in long-term predictions.
14 sharesSource ↗
The study uses machine learning to identify factors affecting AI readiness in businesses across 40 nations, finding that AI readiness is more predictable in developing countries, with scientific research output, internet infrastructure, and public consumption expense as key factors.
13 sharesSource ↗
The research uses sentiment analysis and machine learning to predict stock indexes, highlighting the impact of investor sentiment and exchange rates on the Shanghai Composite Index.
14 sharesSource ↗
Machine Learning methods are more effective than Classical Bayesian methods in predicting unit root in univariate time-series models, particularly when there is class imbalance.
14 sharesSource ↗
The study presents the sequential random search (SQRS) for hyperparameter tuning in machine learning, which reduces computational effort by discarding inferior parameter configurations early.
13 sharesSource ↗
The study uses machine learning to predict the results of standard battles in the Chinese solid-state lighting industry, indicating that strong alliances, patent application experience, and marketization level increase a firm's chances of winning.
12 sharesSource ↗
Machine learning can forecast business trends using big data, but its integration necessitates major system changes, including data collection and workflow architecture.
52 sharesSource ↗
The study reveals that individual investor attention can solve the volatility puzzle in China's stock market, and securities analysts can lessen market information asymmetry.
16 sharesSource ↗
Machine learning models using extensive financial data can predict future earnings changes more accurately than traditional models and professional analysts.
93 sharesSource ↗
A machine learning system can identify financial misinformation on social media based on the truth-default theory, impacting both theory and practice.
32 sharesSource ↗
Bid-rigging cartels can be identified using statistical screening methods and machine learning, but their effectiveness varies across countries due to institutional differences.
24 sharesSource ↗
The research shows that the returns of portfolios based on long-short anomalies can predict overall market returns, due to the persistence of overpricing correction and asymmetric limits of arbitrage.
116 sharesSource ↗
The paper proposes a factor timing strategy using deep learning and 146 characteristic-based factors, which performs better than other portfolios, especially in the Chinese stock market.
52 sharesSource ↗
The research indicates that mixed-frequency factor models provide better forecasts of the Chinese economy, although they were not significantly superior during the Global Financial Crisis.
10 sharesSource ↗
The paper introduces a framework for handling uncertainty in factor models, demonstrating that integrated models perform well out-of-sample and that model uncertainty increases perceived risk in stocks.
13 sharesSource ↗
The study uses various signals to build an efficient currency portfolio, finding that a majority of risk in these assets does not impact their risk premiums.
21 sharesSource ↗
The general machine-learning papers the letter carried in 2023-25.
20 items
A new online learning method for AI training on large medical data sets has been proposed, improving efficiency and reducing data loss, with a 15% improvement in multi-organ and tumor segmentation.
14 shares11 citations todaySource ↗
An AI-guided infrastructure is being developed to better understand and apply materials informatics, with potential applications in designing materials for gas turbines, jet engines, and hypersonic vehicles.
11 shares1 citation todaySource ↗
A new benchmark has been proposed to evaluate the reliability of large vision-language models as visual aids for visually impaired individuals in diverse cultural settings.
9 shares14 citations todaySource ↗
Lazarus, a system for training Sparsely-activated Mixture-of-Experts models, has been introduced, which improves workload balance and recovery probability, outperforming existing systems in frequent node failures and real spot instance trace.
7 shares13 citations todaySource ↗
A new model, Joint-Image Diffusion, has been proposed for personalized text-to-image generation, which learns the joint distribution of multiple related text-image pairs, outperforming previous personalization models.
7 shares66 citations todaySource ↗
The article introduces a new efficient algorithm for Betti matching calculation, which enhances the training of topology aware segmentation networks and speeds up the process.
6 shares7 citations todaySource ↗
HDR Radiance Acquisition: The paper discusses PanDORA, a system using two 360-degree cameras to capture indoor scenes in high dynamic range, improving HDR radiance reconstruction without compromising visual quality.
6 shares1 citation todaySource ↗
Representation Learning for Electrolyte Design: The authors present Uni-ELF, a representation learning framework for electrolyte design, which surpasses existing methods in predicting molecular and formulation properties and can be incorporated into an automatic experimental design process.
5 shares3 citations todaySource ↗
The paper explores the link between learning rate tuning and curvature in machine learning, presenting a new method, Curvature Dynamics Aware Tuning (CDAT), which focuses on long-term curvature stabilization.
4 shares9 citations todaySource ↗
The article introduces Xiangqi Structurally Variable (XQSV), a deep learning architecture that mimics human behavior in Chinese Chess, achieving around 40% predictive accuracy and passing a three-terminal Turing Test.
4 shares1 citation todaySource ↗
Simple Preference Optimization: SimPO improves reinforcement learning from human feedback by using the average log probability of a sequence as the implicit reward, enhancing training stability and computational efficiency.
728 shares1,173 citations todaySource ↗
A new circuit breaker approach can interrupt harmful AI outputs, providing a significant advancement in protection against harmful behavior and adversarial attacks.
497 shares345 citations todaySource ↗
A Versatile Large Vision Language Model: InternLM-XComposer-2.5, a large-vision language model, excels in text-image comprehension and composition, outperforming existing models on 16 benchmarks.
204 shares210 citations todaySource ↗
The found-in-the-middle calibration mechanism addresses the lost-in-the-middle problem in large language models, improving the model's ability to locate relevant information and enhancing performance.
184 shares131 citations todaySource ↗
Enhancing Diffusion Models: Discrete-Continuous Latent Variable Diffusion Models simplify encoding complex data into a Gaussian distribution, improving performance in image synthesis and molecular docking tasks.
73 shares24 citations todaySource ↗
The article explores how latent diffusion models can recreate complex music from brainwave (EEG) recordings, aiding in brain-computer interface research.
38 shares9 citations todaySource ↗
Cross-Lingual Similarity: The study introduces mPLMSim, a language similarity measure tool that enhances cross-lingual transfer performance by 1%-2% using multilingual pretrained language models.
25 shares1 citation todaySource ↗
Foresight Minds: The paper presents the integration of future modeling into Multimodal Large Language Models (MLLMs) to improve their predictive capabilities, resulting in a new MLLM called Merlin.
16 shares52 citations todaySource ↗
The research proposes a new learning paradigm for continuous and unsupervised class discovery in class-iNCD, using self-supervised pre-trained models.
13 shares15 citations todaySource ↗
The article introduces ByITFL, a new Federated Learning scheme that protects against malicious users and ensures data privacy, marking the first Byzantine resilient scheme with full information-theoretic privacy.
11 shares18 citations todaySource ↗
Papers that shipped their code, from the Papers with Code feed (2023-25).
13 items
IXC2.5's longcontext feature allows it to excel in tasks that need large amounts of input and output contexts.
2,079 shares
KVCachecentric: Mooncake surpasses the standard method by up to 525% in data processing speed in specific simulated situations, while still meeting service level objectives.
773 shares
Animation: The research investigates the implicit-keypoint-based framework's potential to efficiently manage computational resources and control, deviating from common diffusion-based methods.
773 shares
The article explores the trade-off between performance and cost in the selection of large language models.
642 shares
The article presents BM25S, a Python version of BM25 that exclusively uses Numpy and Scipy.
601 shares
The article suggests a novel data synthesis approach using various perspectives in a large language model to create diverse synthetic data.
338 shares
The article explores the intricacies and constraints of existing agent-based methods in software development.
287 shares
The paper presents Diffusion Forcing, a novel training method for improving the clarity of tokens in a diffusion model.
258 shares
The article assesses the efficiency of instantiations ranging from 125M to 1.3B parameters, and compares it with a Transformer and Mamba, a contemporary RNN.
149 shares
The article provides a comprehensive derivation of model predictive path integral control (MPPI) for a unified control sequence distribution.
132 shares
The article explores the use of pattern and sparse indices to enhance the efficiency and reduce latency in long-context language model's prefilling stage.
113 shares
The article suggests a multisegment training method for ConsistencyFM to strike a balance between the quality of sampling and speed.
64 shares
The article emphasizes the superior performance and flexibility of the point-based representation OSP in comparison to other existing methods.
27 shares
Repositories the letter featured.
10 items
The article talks about a Python-based trading platform for Interactive Brokers.
60 shares
The piece examines a Python framework for algorithmic trading in the cryptocurrency market.
462 shares
The article explains a Python version of the introduction to reinforcement learning.
13,332 shares
The piece discusses the creation of options trading strategies using technical and quantitative methods.
730 shares
The article introduces a high-frequency trading model using the Interactive Brokers API, implemented in Python.
2,443 shares
The article offers a tutorial on developing, executing, and enhancing machine learning applications for production.
36,579 shares
The article explores an unofficial Python client for the Uniswap cryptocurrency exchange.
906 shares
The article details the development of a universal vector search SQLite extension.
973 shares
The article presents a Python-powered web scraper that uses artificial intelligence.
12,904 shares
The article introduces OpenCodeInterpreter, an open-source system aimed at improving code generation through execution and iterative refinement.
1,492 shares
Industry news: funds, hiring, markets and regulation.
19 items
Macrobond is using alternative data and machine learning to enhance economic forecasting and quantitative trading strategies.
16 shares
The article discusses the debate between using traditional methods or AI for decision making.
6 shares
Alphabeth, a hedge fund focused on digital assets, has won several awards at the Hedgeweek Global Digital Assets Awards 2024 for its unique investment models.
5 shares
BitMEX has launched the MEMEMEXTUSDT Basket Index Perpetual Swap Contract, enabling traders to invest in the top 10 meme coins via a single contract.
4 shares
The Market Misconduct Tribunal in Hong Kong has ordered Jonathan Dominic Iu Wai Ching to return over 5.6m from illegal profits made from false trading at Tarascon Capital Management.
4 shares
Increased regulation of hedge fund strategies by Asian authorities may reduce trading activity and affect market liquidity.
3 shares
Yacine Bourezak has been appointed as Senior Analyst at London-based investment firm, Shiprock Capital Management.
3 shares
Jain Global, a new hedge fund firm, has recruited Ali Rauf from Squarepoint Capital to be a portfolio manager.
3 shares
Australia's $2.6tn pension market is cutting back on hedge fund investments due to increased regulatory scrutiny on fees and performance.
3 shares
EEX Group's global power markets experienced a significant rise in monthly volume growth in June, especially in Belgian and CSEE power futures.
2 shares
Major hedge funds like Point72, Citadel, and Millennium Management have bounced back strongly in the first half of 2024 after a poor performance in 2023.
2 shares
Hedge funds have placed high bets on the pound strengthening, a six-year record, due to confidence in a Labour win in the upcoming UK election.
2 shares
Ex-Goldman Sachs analyst, Murad Mahmudov, has criticized Ripple's XRP and Cardano's ADA, labeling them as cult meme coins.
2 shares
Elliott Investment Management has intensified its campaign against Southwest Airlines, criticizing the board for implementing a poison pill strategy and appointing a new director.
2 shares
CoinShares reports that digital asset funds saw a $441m inflow last week, likely due to buying opportunities following price drops from Mt Gox and the German Government's selling pressure.
2 shares
Avelacom has joined the RCB Rede de Comunicação B3 B3 Communications Network, enhancing its low latency solution offerings for the Brazil Stock Exchange.
1 shares
Financial markets communication provider IPC has been certified as a Google Cloud Interconnect Partner, expanding its global Supported Services.
1 shares
Zurich-based asset management group GAM has appointed Randel Freeman as Co-Head/Co-CIO of its new unit, GAM Alternatives.
1 shares
Westpac's new international technology head is returning from a stint at a major tech company.
1 shares
Episodes on markets, quant methods and economics.
8 items
David Hand's book Dark Data: Why What You Don't Know Matters emphasizes the significance of uncollected or unanalyzed data in data analysis.
12 shares
Jerry Parker, CEO of Chesapeake Holding Company, announces the launch of a new ETF, the Cambria Chesapeake Pure Trend ETF, that follows a systematic trend across various assets.
9 shares
Despite slowing macroeconomic growth, China's commodity imports are at a record high, with the country implementing a three-pronged strategy for long-term security of strategic commodities.
8 shares
JPMorgan strategists share their predictions on the Treasury Inflation, Short Term, and Derivatives markets for the rest of the year.
6 shares
Tyler Wood examines the technical health of US equity markets, suggesting potential growth in cruise liners, biotech healthcare, and news media sectors.
3 shares
A podcast with Sonja Laud, James Carrick, and Chris Jeffery discusses how geopolitics, immigration, deglobalisation, and fiscal versus monetary policy affect investors, stressing the need for a diverse asset portfolio.
3 shares
Daan Struyven and Nikhil Bhandari from Goldman Sachs Research discuss the slowed transition from oil due to increased incomes and lower electric vehicle sales, and its implications for investors and consumers.
0 shares
Henry Occelston's podcast features Shahin Vallée discussing the potential outcomes of Macron's snap election in France, including the possibility of a far-right majority.
0 shares
Posts from quant and economics blogs and newsletters.
4 items
The Relative Strength Index (RSI) is a technical analysis tool used to measure price momentum and change.
3 shares
The Stoller analysis tool is frequently used by traders to predict market movements due to its simplicity.
0 shares
The author reminisces about their influential years in Britain, highlighting its impact on their personality and concern for the country's future.
0 shares
The second article explores the process and consequences of arriving at a negative decision or conclusion.
0 shares
Posts from quant researchers on X.
6 items
The study reveals that global currency volatility and variance metrics can predict currency returns for 3 to 12 months.
3 shares
Dotsis' research introduces a novel option measure that detects anomalies in risk-neutral distribution, greatly impacting the SP 500 asset allocation strategy.
2 shares
A new book discussing the application of Generative AI and ChatGPT in financial forecasting has been released.
1 shares
The article provides an in-depth analysis of the low-risk anomaly across various asset classes and explores possible reasons for its existence.
1 shares
The publication delves into multiasset carry strategies, discussing its history, theory, benchmarks, related literature, and its application in stacking.
0 shares
The piece investigates the idea of mental mapping and representation, using the sound of a duck's quack as a metaphor.
0 shares