{"id":8577,"date":"2022-03-09T14:32:00","date_gmt":"2022-03-09T14:32:00","guid":{"rendered":"https:\/\/www.quoniam.com\/?p=8577"},"modified":"2023-11-29T09:30:25","modified_gmt":"2023-11-29T09:30:25","slug":"artificial-intelligence-at-quoniam","status":"publish","type":"post","link":"https:\/\/www.quoniam.com\/en\/article\/artificial-intelligence-at-quoniam\/","title":{"rendered":"Artificial intelligence at Quoniam"},"content":{"rendered":"\n
\n

Terms like artificial intelligence (AI), machine learning (ML) and natural language processing (NLP) are all over the press. Quoniam started analysing information with machine-learning methods in 2016. Since 2018, they have been part of the investment process. But how do these methods add value in live portfolios?<\/p>\n\n\n\n

As an active quantitative asset manager, we generate added value for investors by processing a wide range of information into superior return and risk forecasts and constructing portfolios on this basis. In doing so, we are confronted with three challenges.<\/p>\n<\/div><\/div>\n\n\n\n

\n
Challenges in quantitative asset management<\/h5>\n\n\n\n
    \n
  1. In competitive financial markets, easily processable information is often already priced in.<\/li>\n\n\n\n
  2. The amount of potentially relevant information is huge and growing faster and faster.<\/li>\n\n\n\n
  3. Unstructured data such as text or images require new techniques.<\/li>\n<\/ol>\n\n\n\n

    Methods such as machine learning and natural language processing help us to overcome these challenges. We explain how this works in the following.<\/p>\n<\/div><\/div>\n\n\n\n

    \n
    Modelling of non-linear relationships <\/h5>\n\n\n\n

    What information is hidden in the data? Which correlations are valuable in order to better predict the return and risk of a security? These are the kinds of questions we address when we apply machine learning in research. This way, we can capture correlations that are not visible at first glance. <\/p>\n\n\n\n

    Financial markets, for example, exhibit numerous non-linear relationships. For example, stocks with low-dividend yields tend to underperform stocks with high-dividend yields. However, companies that do not pay dividends are often growth companies that outperform strongly. A linear relationship that assumes a high-dividend yield is better than a low one cannot reflect this. Machine learning allows us to more easily grasp such complex relationships.<\/p>\n<\/div><\/div>\n\n\n\n

    \n\n\n
    \n
    \n
    \n

    \u2018Machine learning is an important tool to generate excess returns over widely known smart-beta premiums. We have defined clear rules for the flow of research projects and implemented large parts of our ML pipeline ourselves to ensure economically relevant results in aspects like cross validation or handling holdout data.\u2019<\/p>\n\n <\/div>\n <\/div>\n

    \n
    \n \"\"\n <\/div>\n
    \n

    Dr Maximilian Stroh<\/strong>, CFA
    \nHead of Research<\/p>\n\n <\/div>\n <\/div>\n<\/div>\n<\/div><\/div>\n\n\n\n

    <\/div>\n\n\n\n
    \n

    In doing so, we attach great importance to the interaction between man and machine. A purely technology-based application of machine learning in asset management carries the risk of depicting spurious correlations. That is why it is important that humans decide which data is considered and what the fundamental structure of the return forecast model is. With the help of human plausibility checks, machine learning can also capture complex correlations and optimise investment strategies on this basis.<\/p>\n<\/div><\/div>\n\n\n\n

    \n
    Efficient handling of data records<\/h5>\n\n\n\n

    The range of available data sets in asset management has exploded in recent years. It has long since ceased to be possible to manually analyse every potentially relevant data set. This is one of the reasons why the integration of machine learning into Quoniam\u2019s research processes is progressing continuously.<\/p>\n<\/div><\/div>\n\n\n\n

    \n\n\n
    \n
    \n
    \n

    \u2018By using machine learning, we can examine new data sets for interesting correlations and generate ideas for research projects much faster than before. ML makes it easier to focus on the most promising data sets and analyse them in detail.\u2019<\/p>\n\n <\/div>\n <\/div>\n

    \n
    \n \"\"\n <\/div>\n
    \n

    Stefan Klein<\/strong>, CFA
    \nResearch Forecasts<\/p>\n\n <\/div>\n <\/div>\n<\/div>\n<\/div><\/div>\n\n\n\n

    <\/div>\n\n\n\n
    \n

    Machine learning is excellently suited, for example, to obtaining an initial assessment of high-dimensional data sets with many potential factor candidates as to whether an additional benefit can be achieved with the data set compared to existing forecast models \u2013 and this with little manual effort. This highly efficient filtering to the data sets most relevant for our investment process subsequently gives us the time to analyse the best candidates in depth for plausible economic correlations and robustness.<\/p>\n\n\n\n

    In addition to using modern algorithms, we are also continuously optimising our research platform<\/a> and the technical infrastructure behind it to be able to process even more data in even less time.<\/p>\n<\/div><\/div>\n\n\n\n

    \n
    Analysis of texts, images and other high-dimensional data<\/h5>\n\n\n\n

    We live in a time of information overload and big data. Unstructured data such as text or image data is gaining in importance. Here, too, machine learning and artificial intelligence come into play: such unstructured data sets can often only be evaluated and structured using AI.<\/p>\n<\/div><\/div>\n\n\n\n

    \n\n\n
    \n
    \n
    \n

    \u2018Bringing a machine to the mental level of a human and imitating their creativity is hardly possible in asset management. But we can use modern technologies to systematically evaluate unstructured data and extract valuable signals from it that we can use to improve our forecasts.\u2019<\/p>\n\n <\/div>\n <\/div>\n

    \n
    \n \"\"\n <\/div>\n
    \n

    Dr Volker Fl\u00f6gel<\/strong>, CFA
    \nChief Investment Officer<\/span><\/p>\n\n <\/div>\n <\/div>\n<\/div>\n<\/div><\/div>\n\n\n\n

    <\/div>\n\n\n\n
    \n

    One example is natural language processing (NLP)<\/strong>, which refers to the algorithmic evaluation of text data. For example, we analyse annual and quarterly reports of US companies for the aspects of sentiment and tone. Are they formulated in a light or complex way? Have certain sections changed compared to the previous year? The evaluation yields important insights for our return forecast models.<\/p>\n<\/div><\/div>\n\n\n\n

    \n\n\n
    \n
    \n
    Conclusion: artificial intelligence is an indispensable tool for modern asset management<\/div>\n
    \n

    Machine-learning models capture complex relationships between variables.<\/p>\n\n <\/div>\n \n\n <\/div>\n

    \n
    Conclusion: artificial intelligence is an indispensable tool for modern asset management<\/div>\n
    \n

    Research process is becoming more efficient and more research projects can be realised in a short period of time.<\/p>\n\n <\/div>\n \n\n <\/div>\n

    \n
    Conclusion: artificial intelligence is an indispensable tool for modern asset management<\/div>\n
    \n

    Analysis of texts, images and other high-dimensional data \u2013 unstructured data is becoming structured data.<\/p>\n\n <\/div>\n \n\n <\/div>\n <\/div>\n<\/div><\/div>\n\n\n\n

    THIS MIGHT ALSO BE INTERESTING TO YOU<\/h6>\n\n\n\n
    \n\n\n
    \n
    \n \n
    \n \n
    \n \n
    \n
    \n \n <\/div>\n
    \n
    \n
    \n Interview\n <\/div>\n
    \n October 2021\n <\/div>\n <\/div>\n
    Cloud, Python & Co.: next-level research at Quoniam<\/div>\n
    \n

    Setting up efficient research in the cloud, migrating tens of thousands of lines of code to Python \u2013 and doing it all largely from home? Our research team has made it possible. An interview with Dr Volker Fl\u00f6gel, Head of Research, and Andre Fr\u00f6hlich, Head of Research Technology.<\/p>\n\n <\/div>\n <\/div>\n <\/div>\n <\/a>\n <\/div>\n <\/div>\n \n

    \n \n
    \n \n
    \n
    \n \n <\/div>\n
    \n
    \n
    \n Interview\n <\/div>\n
    \n February 2022\n <\/div>\n <\/div>\n
    Expert discussion on the use of Big Data: “We want to keep the information edge on unstructured data”<\/div>\n
    \n

    In conversation with Markus Ebner, Head of Multi-Asset at Quoniam Asset Management.<\/p>\n\n <\/div>\n <\/div>\n <\/div>\n <\/a>\n <\/div>\n <\/div>\n \n

    \n \n
    \n \n
    \n
    \n \n <\/div>\n
    \n
    \n
    \n Article\n <\/div>\n
    \n February 2022\n <\/div>\n <\/div>\n
    Making investment decisions with alternative data<\/div>\n
    \n

    How can asset managers gain an information advantage in today’s world? One way is to use unstructured data. They provide valuable signals and can thus offer an important investment advantage.<\/p>\n\n <\/div>\n <\/div>\n <\/div>\n <\/a>\n <\/div>\n <\/div>\n <\/div>\n <\/div>\n<\/div><\/div>\n","protected":false},"excerpt":{"rendered":"

    Terms like artificial intelligence (AI), machine learning (ML) and natural language processing (NLP) are all over the press. Quoniam started analysing information with machine-learning methods in 2016. Since 2018, they have been part of the investment process. But how do these methods add value in live portfolios? As an active quantitative asset manager, we generate […]<\/p>\n","protected":false},"author":12,"featured_media":13547,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"_seopress_robots_primary_cat":"none","_seopress_titles_title":"Artificial intelligence and institutional investments","_seopress_titles_desc":"How can artificial intelligence and machine learning improve institutional investments? Learn more!","_seopress_robots_index":"","footnotes":""},"categories":[44],"tags":[89,64,105,55],"class_list":["post-8577","post","type-post","status-publish","format-standard","has-post-thumbnail","category-article","tag-data","tag-factor-investing","tag-machine-learning","tag-technology"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.quoniam.com\/en\/wp-json\/wp\/v2\/posts\/8577","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.quoniam.com\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.quoniam.com\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.quoniam.com\/en\/wp-json\/wp\/v2\/users\/12"}],"replies":[{"embeddable":true,"href":"https:\/\/www.quoniam.com\/en\/wp-json\/wp\/v2\/comments?post=8577"}],"version-history":[{"count":60,"href":"https:\/\/www.quoniam.com\/en\/wp-json\/wp\/v2\/posts\/8577\/revisions"}],"predecessor-version":[{"id":132818,"href":"https:\/\/www.quoniam.com\/en\/wp-json\/wp\/v2\/posts\/8577\/revisions\/132818"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.quoniam.com\/en\/wp-json\/wp\/v2\/media\/13547"}],"wp:attachment":[{"href":"https:\/\/www.quoniam.com\/en\/wp-json\/wp\/v2\/media?parent=8577"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.quoniam.com\/en\/wp-json\/wp\/v2\/categories?post=8577"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.quoniam.com\/en\/wp-json\/wp\/v2\/tags?post=8577"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}