We have taken the application of AI in finance to another level

Distinguishing information from noise, analysing and making sense out of an unmanageable amount of structured and unstructured data available today is beyond human capacity.

With the ever-increasing application and popularity of natural language processing in the context of business, Suzugia is taking the next step, pushing the boundaries of these common applications to augment the capability of investors, analysts and researchers by continuously and tirelessly exploring, screening, learning and applying the learnings to find non-intuitive inferences and gives them an unparalleled competitive edge.

  • Beyond Sentiment

Suzugia does not only scan and analyse deeply and broadly, it guides you through your decision-making process by identifying cause-and-effect relationships amongst a chain of factors and events, which are way beyond sentiment. Causality analysis enables longer-lasting insight and forward-looking decisions.

  • Proprietary AI

Our novel optimisation method enables us to do more with less. We have invented a technique that optimises computational processing time and memory. We use distributional semantics combined with advanced non-smooth optimisation algorithms at the core of the unsupervised deep learning stage to produce rich and sparse coding of any language.

Read More

We also create representations and semantic networks based on semi-supervised learning to establish paradigmatic relations like synonymy, hyponymy, antonymy, and entailment (deduction, implication, causation, etc) as well as sense distinctions. We leverage and enrich existing lexical database such as Wordnet (Princeton University).

  • Sparsity

The key success factor to make efficient and resilient decisions in a volatile, uncertain, complex and ambiguous world is not that much accessing a large quantity of data  but rather selecting the right information (true information) at the right time; finding the needle in high-dimensional haystack!

Read More

Since the beginning of the 21st century, the size, breadth, and granularity of data is growing exponentially in both economic and scientific sectors. The complexity of such “big data” repositories offer new opportunities and pose new challenges to investigate nonlinear relationships between causes and effects, to explain and to make sense before trying to predict.
Sparsity is at the core of our coding of language as well as our exploration techniques.

  • Explainability & Transparency, not black-boxes

Suzugia aims to augment humans’ process of decision-making and by no means to cut the role of humans out of the process. Therefore explainability and transparency are key preconditions. In recent years, Natural Language Processing (NLP) has found widespread application in the business context.

Read More

A few companies/organizations are distributing open source packages that are widely used to make trading decisions such as Google AI. The risk is that these packages are offered as Black Boxes and the ”why” and the “how” of outcomes is not explained and transparent. Sparsity (focusing on the relevant information only) and navigation through graph-structured data enable us to obtain traceability and explainability of all our computer-made analysis and decision process.

  • Continuous Human-Computer Interactions(HCI) and Collaborations
    (from social listening to reasoning & efficient decision making)

We use graph-structured data and associated mathematical tools (e.g. spectral analysis) to design human-computer interactions to stimulate decision makers at an individual and collective level, creating a “decision driven mirror” proposing factors and features, which are likely to influence the decision process of the group.

Read More

Continuous Human-Computer Interactions (HCI) and Collaborations (from social listening to social reasoning and efficient decision making.

We use graph-structured data and associated mathematical tools (e.g. spectral analysis) to design human-computer interactions to stimulate decision makers at an individual and collective level, creating a “decision driven mirror” proposing factors and features, which are likely to influence the decision process of the group.