Gaurav Oberoi, P. Poduval, Karamjit Singh, Sangam Verma, Pranay Gupta
Similarity-search is an important problem to solve for the payment industry having user-merchant interaction data. It finds out merchants similar to a given merchant and solves various tasks like peer-set generation, recommendation, community detection, and anomaly detection. Recent works have shown that by leveraging interaction data, Graph Neural Networks (GNNs) can be used to generate node embeddings for entities like a merchant, which can be further used for such similarity-search tasks. However, most of the real-world financial data come with high cardinality categorical features such as city, industry, super-industries, etc. which are fed to the GNNs in a one-hot encoded manner. Current GNN algorithms are not designed to work for such sparse features which makes it difficult for them to learn these sparse features preserving embeddings. In this work, we propose CaPE, a Category Preserving Embedding generation method which preserves the high cardinality feature information in the embeddings. We have designed CaPE to preserve other important numerical feature information as well. We compare CaPE with the latest GNN algorithms for embedding generation methods to showcase its superiority in peer set generation tasks on real-world datasets, both external as well as internal (synthetically generated). We also compared our method for a downstream task like link prediction.
{"title":"CaPE: Category Preserving Embeddings for Similarity-Search in Financial Graphs","authors":"Gaurav Oberoi, P. Poduval, Karamjit Singh, Sangam Verma, Pranay Gupta","doi":"10.1145/3533271.3561788","DOIUrl":"https://doi.org/10.1145/3533271.3561788","url":null,"abstract":"Similarity-search is an important problem to solve for the payment industry having user-merchant interaction data. It finds out merchants similar to a given merchant and solves various tasks like peer-set generation, recommendation, community detection, and anomaly detection. Recent works have shown that by leveraging interaction data, Graph Neural Networks (GNNs) can be used to generate node embeddings for entities like a merchant, which can be further used for such similarity-search tasks. However, most of the real-world financial data come with high cardinality categorical features such as city, industry, super-industries, etc. which are fed to the GNNs in a one-hot encoded manner. Current GNN algorithms are not designed to work for such sparse features which makes it difficult for them to learn these sparse features preserving embeddings. In this work, we propose CaPE, a Category Preserving Embedding generation method which preserves the high cardinality feature information in the embeddings. We have designed CaPE to preserve other important numerical feature information as well. We compare CaPE with the latest GNN algorithms for embedding generation methods to showcase its superiority in peer set generation tasks on real-world datasets, both external as well as internal (synthetically generated). We also compared our method for a downstream task like link prediction.","PeriodicalId":134888,"journal":{"name":"Proceedings of the Third ACM International Conference on AI in Finance","volume":"107 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131847889","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Successful predictive models for financial applications often require harnessing complementary information from multiple datasets. Incorporating data from different sources into a single model can be challenging as they vary in structure, dimensions, quality, and completeness. Simply merging those datasets can cause redundancy, discrepancy, and information loss. This paper proposes a convolutional neural network-based nonlinear tensor coupling and completion framework (NLTCC) to combine heterogeneous datasets without compromising data quality. We demonstrate the effectiveness of NLTCC in solving a specific business problem - predicting firms’ earnings from financial analysts’ earnings forecast. First, we apply NLTCC to fuse firm characteristics and stock market information into the financial analysts’ earnings forecasts data to impute missing values and improve data quality. Subsequently, we predict the next quarter’s earnings based on the imputed data. The experiments reveal that the prediction error decreases by 65% compared with the benchmark analysts’ consensus forecast. The long-short portfolio returns based on NLTCC outperform analysts’ consensus forecast and the S&P-500 index from three-day up to two-month holding period. The prediction accuracy improvement is robust with different performance metrics and various industry sectors. Notably, it is more salient for the sectors with higher heterogeneity.
{"title":"Machine Learning for Earnings Prediction: A Nonlinear Tensor Approach for Data Integration and Completion","authors":"Ajim Uddin, Xinyuan Tao, Chia-Ching Chou, Dantong Yu","doi":"10.1145/3533271.3561677","DOIUrl":"https://doi.org/10.1145/3533271.3561677","url":null,"abstract":"Successful predictive models for financial applications often require harnessing complementary information from multiple datasets. Incorporating data from different sources into a single model can be challenging as they vary in structure, dimensions, quality, and completeness. Simply merging those datasets can cause redundancy, discrepancy, and information loss. This paper proposes a convolutional neural network-based nonlinear tensor coupling and completion framework (NLTCC) to combine heterogeneous datasets without compromising data quality. We demonstrate the effectiveness of NLTCC in solving a specific business problem - predicting firms’ earnings from financial analysts’ earnings forecast. First, we apply NLTCC to fuse firm characteristics and stock market information into the financial analysts’ earnings forecasts data to impute missing values and improve data quality. Subsequently, we predict the next quarter’s earnings based on the imputed data. The experiments reveal that the prediction error decreases by 65% compared with the benchmark analysts’ consensus forecast. The long-short portfolio returns based on NLTCC outperform analysts’ consensus forecast and the S&P-500 index from three-day up to two-month holding period. The prediction accuracy improvement is robust with different performance metrics and various industry sectors. Notably, it is more salient for the sectors with higher heterogeneity.","PeriodicalId":134888,"journal":{"name":"Proceedings of the Third ACM International Conference on AI in Finance","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123434035","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
As intelligent trading agents based on reinforcement learning (RL) gain prevalence, it becomes more important to ensure that RL agents obey laws, regulations, and human behavioral expectations. There is substantial literature concerning the aversion of obvious catastrophes like crashing a helicopter or bankrupting a trading account, but little around the avoidance of subtle non-normative behavior for which there are examples, but no programmable definition. Such behavior may violate legal or regulatory, rather than physical or monetary, constraints. In this article, I consider a series of experiments in which an intelligent stock trading agent maximizes profit but may also inadvertently learn to spoof the market in which it participates. I first inject a hand-coded spoofing agent to a multi-agent market simulation and learn to recognize spoofing activity sequences. Then I replace the hand-coded spoofing trader with a simple profit-maximizing RL agent and observe that it independently discovers spoofing as the optimal strategy. Finally, I introduce a method to incorporate the recognizer as normative guide, shaping the agent’s perceived rewards and altering its selected actions. The agent remains profitable while avoiding spoofing behaviors that would result in even higher profit. After presenting the empirical results, I conclude with some recommendations. The method should generalize to the reduction of any unwanted behavior for which a recognizer can be learned.
{"title":"Learning Not to Spoof","authors":"David Byrd","doi":"10.1145/3533271.3561767","DOIUrl":"https://doi.org/10.1145/3533271.3561767","url":null,"abstract":"As intelligent trading agents based on reinforcement learning (RL) gain prevalence, it becomes more important to ensure that RL agents obey laws, regulations, and human behavioral expectations. There is substantial literature concerning the aversion of obvious catastrophes like crashing a helicopter or bankrupting a trading account, but little around the avoidance of subtle non-normative behavior for which there are examples, but no programmable definition. Such behavior may violate legal or regulatory, rather than physical or monetary, constraints. In this article, I consider a series of experiments in which an intelligent stock trading agent maximizes profit but may also inadvertently learn to spoof the market in which it participates. I first inject a hand-coded spoofing agent to a multi-agent market simulation and learn to recognize spoofing activity sequences. Then I replace the hand-coded spoofing trader with a simple profit-maximizing RL agent and observe that it independently discovers spoofing as the optimal strategy. Finally, I introduce a method to incorporate the recognizer as normative guide, shaping the agent’s perceived rewards and altering its selected actions. The agent remains profitable while avoiding spoofing behaviors that would result in even higher profit. After presenting the empirical results, I conclude with some recommendations. The method should generalize to the reduction of any unwanted behavior for which a recognizer can be learned.","PeriodicalId":134888,"journal":{"name":"Proceedings of the Third ACM International Conference on AI in Finance","volume":"104 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131742458","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Increasing access to financial services data helps accelerate the monitoring and management of datasets and facilitates better business decision-making. However, financial services datasets are typically vast, ranging in terabytes of data, containing both structured and unstructured. It is a laborious task to comb through all the data and map them reasonably. Mapping the data is important to perform comprehensive analysis and take informed business decisions. Based on client engagements, we have observed that there is a lack of industry standards for definitions of key terms and a lack of governance for maintaining business processes. This typically leads to disconnected siloed datasets generated from disintegrated systems. To address these challenges, we developed a novel methodology DaME (Data Mapping Engine) that performs data mapping by training a data mapping engine and utilizing human-in-the-loop techniques. The results from the industrial application and evaluation of DaME on a financial services dataset are encouraging that it can help reduce manual effort by automating data mapping and reusing the learning. The accuracy from our dataset in the application is much higher at 69% compared to the existing state-of-the-art with an accuracy of 34%. It has also helped improve the productivity of the industry practitioners, by saving them 14,000 hours of time spent manually mapping vast data stores over a period of ten months.
{"title":"Mapping of Financial Services datasets using Human-in-the-Loop","authors":"Shubhi Asthana, R. Mahindru","doi":"10.1145/3533271.3561705","DOIUrl":"https://doi.org/10.1145/3533271.3561705","url":null,"abstract":"Increasing access to financial services data helps accelerate the monitoring and management of datasets and facilitates better business decision-making. However, financial services datasets are typically vast, ranging in terabytes of data, containing both structured and unstructured. It is a laborious task to comb through all the data and map them reasonably. Mapping the data is important to perform comprehensive analysis and take informed business decisions. Based on client engagements, we have observed that there is a lack of industry standards for definitions of key terms and a lack of governance for maintaining business processes. This typically leads to disconnected siloed datasets generated from disintegrated systems. To address these challenges, we developed a novel methodology DaME (Data Mapping Engine) that performs data mapping by training a data mapping engine and utilizing human-in-the-loop techniques. The results from the industrial application and evaluation of DaME on a financial services dataset are encouraging that it can help reduce manual effort by automating data mapping and reusing the learning. The accuracy from our dataset in the application is much higher at 69% compared to the existing state-of-the-art with an accuracy of 34%. It has also helped improve the productivity of the industry practitioners, by saving them 14,000 hours of time spent manually mapping vast data stores over a period of ten months.","PeriodicalId":134888,"journal":{"name":"Proceedings of the Third ACM International Conference on AI in Finance","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132259584","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In a semi-realistic market simulator, independent reinforcement learning algorithms may facilitate market makers to maintain wide spreads even without communication. This unexpected outcome challenges the current antitrust law framework. We study the effectiveness of maker-taker fee models in preventing cooperation via algorithms. After modeling market making as a repeated general-sum game, we experimentally show that the relation between net transaction costs and maker rebates is not necessarily monotone. Besides an upper bound on taker fees, we may also need a lower bound on maker rebates to destabilize the cooperation. We also consider the taker-maker model and the effects of mid-price volatility, inventory risk, and the number of agents.
{"title":"Can maker-taker fees prevent algorithmic cooperation in market making?","authors":"Bingyan Han","doi":"10.1145/3533271.3561685","DOIUrl":"https://doi.org/10.1145/3533271.3561685","url":null,"abstract":"In a semi-realistic market simulator, independent reinforcement learning algorithms may facilitate market makers to maintain wide spreads even without communication. This unexpected outcome challenges the current antitrust law framework. We study the effectiveness of maker-taker fee models in preventing cooperation via algorithms. After modeling market making as a repeated general-sum game, we experimentally show that the relation between net transaction costs and maker rebates is not necessarily monotone. Besides an upper bound on taker fees, we may also need a lower bound on maker rebates to destabilize the cooperation. We also consider the taker-maker model and the effects of mid-price volatility, inventory risk, and the number of agents.","PeriodicalId":134888,"journal":{"name":"Proceedings of the Third ACM International Conference on AI in Finance","volume":"19 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116668082","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Anti-money laundering (AML) regulations mandate financial institutions to deploy AML systems based on a set of rules that, when triggered, form the basis of a suspicious alert to be assessed by human analysts. Reviewing these cases is a cumbersome and complex task that requires analysts to navigate a large network of financial interactions to validate suspicious movements. Furthermore, these systems have very high false positive rates (estimated to be over 95%). The scarcity of labels hinders the use of alternative systems based on supervised learning, reducing their applicability in real-world applications. In this work we present LaundroGraph, a novel self-supervised graph representation learning approach to encode banking customers and financial transactions into meaningful representations. These representations are used to provide insights to assist the AML reviewing process, such as identifying anomalous movements for a given customer. LaundroGraph represents the underlying network of financial interactions as a customer-transaction bipartite graph and trains a graph neural network on a fully self-supervised link prediction task. We empirically demonstrate that our approach outperforms other strong baselines on self-supervised link prediction using a real-world dataset, improving the best non-graph baseline by 12 p.p. of AUC. The goal is to increase the efficiency of the reviewing process by supplying these AI-powered insights to the analysts upon review. To the best of our knowledge, this is the first fully self-supervised system within the context of AML detection.
{"title":"LaundroGraph: Self-Supervised Graph Representation Learning for Anti-Money Laundering","authors":"Mário Cardoso, Pedro Saleiro, P. Bizarro","doi":"10.1145/3533271.3561727","DOIUrl":"https://doi.org/10.1145/3533271.3561727","url":null,"abstract":"Anti-money laundering (AML) regulations mandate financial institutions to deploy AML systems based on a set of rules that, when triggered, form the basis of a suspicious alert to be assessed by human analysts. Reviewing these cases is a cumbersome and complex task that requires analysts to navigate a large network of financial interactions to validate suspicious movements. Furthermore, these systems have very high false positive rates (estimated to be over 95%). The scarcity of labels hinders the use of alternative systems based on supervised learning, reducing their applicability in real-world applications. In this work we present LaundroGraph, a novel self-supervised graph representation learning approach to encode banking customers and financial transactions into meaningful representations. These representations are used to provide insights to assist the AML reviewing process, such as identifying anomalous movements for a given customer. LaundroGraph represents the underlying network of financial interactions as a customer-transaction bipartite graph and trains a graph neural network on a fully self-supervised link prediction task. We empirically demonstrate that our approach outperforms other strong baselines on self-supervised link prediction using a real-world dataset, improving the best non-graph baseline by 12 p.p. of AUC. The goal is to increase the efficiency of the reviewing process by supplying these AI-powered insights to the analysts upon review. To the best of our knowledge, this is the first fully self-supervised system within the context of AML detection.","PeriodicalId":134888,"journal":{"name":"Proceedings of the Third ACM International Conference on AI in Finance","volume":"20 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134250312","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Andrea Coletta, Aymeric Moulin, Svitlana Vyetrenko, T. Balch
Multi-agent market simulators usually require careful calibration to emulate real markets, which includes the number and the type of agents. Poorly calibrated simulators can lead to misleading conclusions, potentially causing severe loss when employed by investment banks, hedge funds, and traders to study and evaluate trading strategies. In this paper, we propose a world model simulator that accurately emulates a limit order book market – it requires no agent calibration but rather learns the simulated market behavior directly from historical data. Traditional approaches fail short to learn and calibrate trader population, as historical labeled data with details on each individual trader strategy is not publicly available. Our approach proposes to learn a unique "world" agent from historical data. It is intended to emulate the overall trader population, without the need of making assumptions about individual market agent strategies. We implement our world agent simulator models as a Conditional Generative Adversarial Network (CGAN), as well as a mixture of parametric distributions, and we compare our models against previous work. Qualitatively and quantitatively, we show that the proposed approaches consistently outperform previous work, providing more realism and responsiveness.
{"title":"Learning to simulate realistic limit order book markets from data as a World Agent","authors":"Andrea Coletta, Aymeric Moulin, Svitlana Vyetrenko, T. Balch","doi":"10.1145/3533271.3561753","DOIUrl":"https://doi.org/10.1145/3533271.3561753","url":null,"abstract":"Multi-agent market simulators usually require careful calibration to emulate real markets, which includes the number and the type of agents. Poorly calibrated simulators can lead to misleading conclusions, potentially causing severe loss when employed by investment banks, hedge funds, and traders to study and evaluate trading strategies. In this paper, we propose a world model simulator that accurately emulates a limit order book market – it requires no agent calibration but rather learns the simulated market behavior directly from historical data. Traditional approaches fail short to learn and calibrate trader population, as historical labeled data with details on each individual trader strategy is not publicly available. Our approach proposes to learn a unique \"world\" agent from historical data. It is intended to emulate the overall trader population, without the need of making assumptions about individual market agent strategies. We implement our world agent simulator models as a Conditional Generative Adversarial Network (CGAN), as well as a mixture of parametric distributions, and we compare our models against previous work. Qualitatively and quantitatively, we show that the proposed approaches consistently outperform previous work, providing more realism and responsiveness.","PeriodicalId":134888,"journal":{"name":"Proceedings of the Third ACM International Conference on AI in Finance","volume":"36 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124779981","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Neural style transfer is a powerful computer vision technique that can incorporate the artistic “style" of one image to the “content" of another. The underlying theory behind the approach relies on the assumption that the style of an image is represented by the Gram matrix of its features, which is typically extracted from pre-trained convolutional neural networks (e.g., VGG-19). This idea does not straightforwardly extend to time series stylization since notions of style for two-dimensional images are not analogous to notions of style for one-dimensional time series. In this work, a novel formulation of time series style transfer is proposed for the purpose of synthetic data generation and enhancement. We introduce the concept of stylized features for time series, which is directly related to the time series realism properties, and propose a novel stylization algorithm, called StyleTime, that uses explicit feature extraction techniques to combine the underlying content (trend) of one time series with the style (distributional properties) of another. Further, we discuss evaluation metrics, and compare our work to existing state-of-the-art time series generation and augmentation schemes. To validate the effectiveness of our methods, we use stylized synthetic data as a means for data augmentation to improve the performance of recurrent neural network models on several forecasting tasks.
{"title":"StyleTime: Style Transfer for Synthetic Time Series Generation","authors":"Yousef El-Laham, Svitlana Vyetrenko","doi":"10.1145/3533271.3561772","DOIUrl":"https://doi.org/10.1145/3533271.3561772","url":null,"abstract":"Neural style transfer is a powerful computer vision technique that can incorporate the artistic “style\" of one image to the “content\" of another. The underlying theory behind the approach relies on the assumption that the style of an image is represented by the Gram matrix of its features, which is typically extracted from pre-trained convolutional neural networks (e.g., VGG-19). This idea does not straightforwardly extend to time series stylization since notions of style for two-dimensional images are not analogous to notions of style for one-dimensional time series. In this work, a novel formulation of time series style transfer is proposed for the purpose of synthetic data generation and enhancement. We introduce the concept of stylized features for time series, which is directly related to the time series realism properties, and propose a novel stylization algorithm, called StyleTime, that uses explicit feature extraction techniques to combine the underlying content (trend) of one time series with the style (distributional properties) of another. Further, we discuss evaluation metrics, and compare our work to existing state-of-the-art time series generation and augmentation schemes. To validate the effectiveness of our methods, we use stylized synthetic data as a means for data augmentation to improve the performance of recurrent neural network models on several forecasting tasks.","PeriodicalId":134888,"journal":{"name":"Proceedings of the Third ACM International Conference on AI in Finance","volume":"157 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116462264","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Kshama Dwarakanath, Danial Dervovic, P. Tavallali, Svitlana Vyetrenko, T. Balch
We propose a novel group of Gaussian Process based algorithms for fast approximate optimal stopping of time series with specific applications to financial markets. We show that structural properties commonly exhibited by financial time series (e.g., the tendency to mean-revert) allow the use of Gaussian and Deep Gaussian Process models that further enable us to analytically evaluate optimal stopping value functions and policies. We additionally quantify uncertainty in the value function by propagating the price model through the optimal stopping analysis. We compare and contrast our proposed methods against a sampling-based method, as well as a deep learning based benchmark that is currently considered the state-of-the-art in the literature. We show that our family of algorithms outperforms benchmarks on three historical time series datasets that include intra-day and end-of-day equity asset prices as well as the daily US treasury yield curve rates.
{"title":"Optimal Stopping with Gaussian Processes","authors":"Kshama Dwarakanath, Danial Dervovic, P. Tavallali, Svitlana Vyetrenko, T. Balch","doi":"10.1145/3533271.3561670","DOIUrl":"https://doi.org/10.1145/3533271.3561670","url":null,"abstract":"We propose a novel group of Gaussian Process based algorithms for fast approximate optimal stopping of time series with specific applications to financial markets. We show that structural properties commonly exhibited by financial time series (e.g., the tendency to mean-revert) allow the use of Gaussian and Deep Gaussian Process models that further enable us to analytically evaluate optimal stopping value functions and policies. We additionally quantify uncertainty in the value function by propagating the price model through the optimal stopping analysis. We compare and contrast our proposed methods against a sampling-based method, as well as a deep learning based benchmark that is currently considered the state-of-the-art in the literature. We show that our family of algorithms outperforms benchmarks on three historical time series datasets that include intra-day and end-of-day equity asset prices as well as the daily US treasury yield curve rates.","PeriodicalId":134888,"journal":{"name":"Proceedings of the Third ACM International Conference on AI in Finance","volume":"99 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132397336","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
We consider a trading marketplace that is populated by traders with diverse trading strategies and objectives. The marketplace allows the suppliers to list their goods and facilitates matching between buyers and sellers. In return, such a marketplace typically charges fees for facilitating trade. The goal of this work is to design a dynamic fee schedule for the marketplace that is equitable and profitable to all traders while being profitable to the marketplace at the same time (from charging fees). Since the traders adapt their strategies to the fee schedule, we present a reinforcement learning framework for simultaneously learning a marketplace fee schedule and trading strategies that adapt to this fee schedule using a weighted optimization objective of profits and equitability. We illustrate the use of the proposed approach in detail on a simulated stock exchange with different types of investors, specifically market makers and consumer investors. As we vary the equitability weights across different investor classes, we see that the learnt exchange fee schedule starts favoring the class of investors with the highest weight. We further discuss the observed insights from the simulated stock exchange in light of the general framework of equitable marketplace mechanism design.
{"title":"Equitable Marketplace Mechanism Design","authors":"Kshama Dwarakanath, Svitlana Vyetrenko, T. Balch","doi":"10.1145/3533271.3561673","DOIUrl":"https://doi.org/10.1145/3533271.3561673","url":null,"abstract":"We consider a trading marketplace that is populated by traders with diverse trading strategies and objectives. The marketplace allows the suppliers to list their goods and facilitates matching between buyers and sellers. In return, such a marketplace typically charges fees for facilitating trade. The goal of this work is to design a dynamic fee schedule for the marketplace that is equitable and profitable to all traders while being profitable to the marketplace at the same time (from charging fees). Since the traders adapt their strategies to the fee schedule, we present a reinforcement learning framework for simultaneously learning a marketplace fee schedule and trading strategies that adapt to this fee schedule using a weighted optimization objective of profits and equitability. We illustrate the use of the proposed approach in detail on a simulated stock exchange with different types of investors, specifically market makers and consumer investors. As we vary the equitability weights across different investor classes, we see that the learnt exchange fee schedule starts favoring the class of investors with the highest weight. We further discuss the observed insights from the simulated stock exchange in light of the general framework of equitable marketplace mechanism design.","PeriodicalId":134888,"journal":{"name":"Proceedings of the Third ACM International Conference on AI in Finance","volume":"31 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128599022","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：