Causal AI “xCausal™” collaborates with causal chain technology developed by the University of Tokyo
Accelerating R&D on causal elucidation in financial and economic domains

VELDT Inc. (Headquartered in Shibuya-ku, Tokyo, CEO Jin Nonogami) has launched a technology collaboration with the Izumi Laboratory of the University of Tokyo for the purpose of deepening the understanding of causal relationships in the financial and economic domains using xCausal™, a Causal AI that allows for quick inference of causal effects from data. Through this collaboration, VELDT has made it possible for xCausal™ to adopt structural information on causal chains extracted from text, such as financial information, by means of a causal chain search technology using natural language processing (causal chain technology) developed by the Izumi Laboratory of the University of Tokyo.

This allows companies and organizations to not only confirm the composition of the causal chain, but also to simulate specific causal effects from the data on “xCausal™”. For example, it is possible to estimate the impact of a specific economic event on a company's business performance and link this to decision-making.

VELDT is expanding its Causal AI platform, xCausal™ to a wide range of domains across industries. The platform allows users to rapidly model cause-and-effect relationships from data and easily incorporate them into digital workflows. Today, many companies and organizations repeatedly formulate and test hypotheses based on correlations in order to understand causes and solve problems. In many such cases, false causal relationships, called spurious correlations, often lead them astray. As a result, inferring causality can be time-consuming and actual causes of a problem are often left unidentified.

xCausal™ is a Causal AI platform based on a Structural Causal Model theory, which can appropriately calculate causal effects from causal structures. It is offered as a SaaS service, along with innovations thoroughly pursued for the ease of use in business settings.

The causal chain technology developed by the Izumi Laboratory of the University of Tokyo extracts and visualizes causal chains from textual information such as economic reports and financial summaries. A starting text can be specified for cause and effect, respectively, and the user can select the recommended causal chain. xCausal™" has the ability to set known causal relationships as rules in advance when creating causal graphs that show the structure of causal relationships. Currently , the information set in the causal chain is made available through an external API of xCausal™. This enables companies using causal chain technology in the financial sector to not only understand the composition of causal relationships, but also to conduct trial and error of causal effects based on data with “xCausal™,” dramatically increasing the efficiency and sophistication of the decision-making process. The use of this functionality will be custom provided through the professional services offered by VELDT.

■ Comment from Professor Kiyoshi Izumi, Department of Systems Innovation, School of Engineering, The University of Tokyo

In recent years, causal AI has attracted significant attention in the field of data analysis in a variety of business practices, including finance. In particular, the report emphasizes the usefulness of integrating both statistical causal inference and narrative causation through natural language processing. The collaborative technology developed this time is expected to significantly enhance the applied value of Causal AI and holds great promise for future advancement.

■ What is xCausal™, developed by the leading Causal AI company VELDT ?

Web site: https://xcausal.com/en/

xCausal™ is a SaaS-based Causal AI platform targeted at business users that can be used without coding, even if you are not a data scientist. It has built-in intelligence technologies that support causal inference, such as Smallytics (causality variable recommendation technology) and CKE-LLM (causal knowledge extraction LLM) developed in-house. Easy to use causal inference based on multiple selectable causal search techniques and structural causal models (SCM) allows to validate hypothetical efficient causal effects on a minute-by-minute basis. It shortens the R&D cycle, improves the success rate of experiments, and enables the productive use of highly skilled human resources. In addition, continuous improvement through AI and reliability assessment technologies will enable systematic utilization of causal models as digital knowledge assets within the organization. VELDT offers professional services* that consistently support these efforts.

※Services include workshops and data consulting on selecting variables necessary for causal modeling, causal modeling and causal inference on behalf of the user, development of a dedicated environment, development of custom functions, implementation of system integration, and development of a usage environment that meets the objectives of corporate users.

According to the latest CAUSAL AI global market research report by MarketsandMarkets, “CAUSAL AI COMPANY EVALUATION REPORT, 2024” VELDT is positioned as a “Progressive” company in the startup sector. The following points were cited as reasons for this evaluation.

• Domain-Agnostic, No-Code Platform That Empowers Business Users
• Full-Stack Causal Intelligence with a Future-Proof Vision
• Built-In Intelligence: Smallytics & CKE-LLM
• Services-Led Delivery for Enterprise-Grade Implementation
• Secure, Compliant, and Configurable for Sensitive Environments
• Design for Interpretability & Control
• Divide-and-Conquer Approach to Complexity
• Dual GTM Strategy with Focused Partner Enablement
• Vertical Diversification with a Common Causal Core

■ About VELDT Inc.

VELDT Inc. is a data science company that develops data analysis technologies to solve social problems and accelerate goodwill innovation. Most mainstream AI today is a black box technology that makes it difficult to explain the basis for calculated outputs. On the other hand, in areas such as human decision making, problem solving, and research and development, a white box approach that can interpret and explain the mechanisms leading to the output is important. We leverage the benefits of both causality-based white-box and correlation-based black-box technologies to provide innovators with “trusted AI.” VELDT creates a positive spiral for human society and the global environment by innovating not just technology, but the way we interact with technology.