Leading Strategies

Transformative Practices

Practical Knowledge

Knowledge is power. With a library of posts, articles, and white papers at your fingertips, you’re prepared to turn every opportunity into a win.


AI in Alternative Investments

Ushanas Shastri, of Viteos Fund Services, highlights the role artificial intelligence has to play in the alternative investments space. He has spent 20 years in building front-to-back office tech solutions for Capital Markets, including high frequency trading, portfolio management and reconciliation systems. He is currently evaluating and adopting Machine Learning within solutions at Viteos Fund Services.

Artificial intelligence (AI) is touted as one of the technological signposts of the future. AI is one of the tech trends expected to be a major disruptor. Along with other tech trends such as big data analytics, it has the potential to transform how hedge funds conduct business.

With their ability to perceive patterns and make predictions, AI is changing the way alternative investment firms operate. The greatest inroads can be seen in non-investment related activities, such as procedural, compliance and statutory operations and reporting.

Open source innovations

Historically, tech developments were confined to labs and shrouded in secrecy until their unveiling. Today, that’s all changed. AI frameworks that help build impactful solutions are available as free and open source software. Google, Linux and Facebook are some examples of the many large tech organisations leading the push for open source innovation.

The Acumos AI Project of Linux, Google’s TensorFlow and Facebook’s FAIR are just a few examples of libraries and platforms that developers are using to build AI solutions. This environment is a true force multiplier. Organisations are speeding up and expanding innovation by leveraging platforms, rather than starting from scratch.

This approach to machine learning is finding its way into alternative investments tech. But it is not being applied across all areas. At present, AI has specific use cases in hedge funds.

AI Implementation: Asset Management Versus Operations

To date, AI has been cautiously embraced on the investment side of the hedge fund industry. Though some hedge funds have deployed AI in innovative business models, AI has not yet been able to replace human expertise in managing assets.

It is still too soon for the newest crop of AI technologies to fully transform the investment landscape. Information about the extent and impact of its use is scarce – a result of limited implementation and confidentiality within the field. With few metrics to evaluate its impact, managers are hesitant to widely adopt AI for the investment process.

A study by KPMG on technology adoption in Alternative Investments pegged the percentage of funds with plans to leverage machine learning in fund management at 19%. This highlights the fact that 81% of funds have no concrete plans in place, at present, on adopting this new technology.

For hedge funds, performance is king. Returns are driven by analytics and efficiency. While on the surface hedge funds appear to be a prime environment for applying machine learning and AI technology, adoption is challenged by compliance and performance pressures.

Two of the biggest hurdles in the adoption of AI for investment strategy are access to quality data and the need for transparency.

Data: Quality and Quantity

Deep learning, a core requirement for true AI, is a type of Machine Learning that uses algorithms to model high-level abstractions from data and perform human-like tasks, such as making predictions or recognising speech and images. To be effective, Deep Learning needs to be fed massive amounts of accurate data. Without volumes of clean data, ML cannot provide highly accurate insights to support decision making.

Most businesses possess pockets of operational data, but the quality of this data is seldom very high, nor is it available in sufficient quantities. In addition, most data requires extensive cleansing to become usable for Machine Learning. Even firms that invest in data cleansing and data management solutions find data quality and quantity an ongoing challenge. Many who are entrenched in ML believe data is more important than the algorithms. Unfortunately, it will be some time before alternative investments data quality and quantity are adequate for AI to take over investment decisions.

Transparency: Performance Versus Clarity

Given regulatory and firm-specific compliance mandates, transparency is a high priority. Yet, in the quest for both performance and transparency, there is a critical trade-off. Algorithms that have greater prediction accuracy are less transparent and difficult to explain. Conversely, algorithms that are easier to explain and are more transparent do not perform as well.

Just as people, especially those at the genius level, have difficulty describing how they derived a result, the same is true for explaining conclusions arrived at via machine learning. Algorithms can be extremely complex and beyond concise articulation. Stakeholders are therefore put in a tough situation. When stakeholders cannot fully understand how an algorithm arrives at results, they are reluctant to place their trust in the final product.

AI Inroads within Hedge Funds

As refinements in AI are made, fintech will meet with greater and greater acceptance. Decision-making will become less dependent on human intelligence and intuition and will move more toward machine learning. Statistics and computing power will come together for unprecedented scale and quality. The result will be insight into unseen patterns and solutions for unique problems.

We are already seeing the applications of AI in operations at some of the most forward thinking hedge funds. Bridgewater Associates is reportedly working on automating the role of management, using insights from algorithmic models built to mimic the brains of the company’s employees in making critical decisions.

That is not to say that ML will totally replace human expertise. Algorithmic innovations will instead supplement the decision making process with data insight. Hedge funds will focus on promoting collaboration between humans and AI in areas where extensive and clean data is available. Reconciliation, which is already heavily automated and tool driven, is one area where AI’s impact is being felt.

Hedge fund performance is the confluence of many areas. From the outside, people attribute it solely to investment strategy. But there’s a complex mid and back-office support mechanism that exerts power. Transforming the complex to the simple is where AI can give firms a competitive edge.

Knowing how and where to apply AI takes judicious evaluation. It requires a smart platform along with expert knowledge of operational processes and data sources, quality and quantity.

AI can augment human decision-making in operations to save time and expense, and minimise error. To learn more about application of Artificial Intelligence in alternative investments, and how it can improve mid and back-office operations, reach out to the team at Viteos at reply@viteos.com.

Leave a reply