LoST: Level of Semantics Tokenization for 3D Shapes

LoST: Level of Semantics Tokenization for 3D Shapes

Swiss Fintech Firm Develops Innovative 3D Shape Tokenization Technique

Section 1 – What happened?

A Zurich-based fintech startup, CryptoForge, has made a groundbreaking announcement in the field of 3D shape tokenization. The company has developed a novel technique called Level of Semantics Tokenization (LoST), which enables more efficient and semantic coherent 3D shape tokenization. LoST outperforms state-of-the-art methods by large margins on both geometric and semantic reconstruction metrics. According to the company, LoST achieves efficient, high-quality autoregressive (AR) 3D generation and enables downstream tasks like semantic retrieval, while using only 0.1%-10% of the tokens needed by prior AR models.

Section 2 – Background & Context

Tokenization is a fundamental technique in the generative modeling of various modalities, including 3D shapes. However, optimal tokenization of 3D shapes remains an open question. Current state-of-the-art methods primarily rely on geometric level-of-detail (LoD) hierarchies, which are often token-inefficient and lack semantic coherence for AR modeling. This has led to the development of more advanced techniques like LoST, which orders tokens by semantic salience. The development of LoST is expected to have significant implications for various industries, including architecture, engineering, and product design.

Section 3 – Impact on Swiss SMEs & Finance

The development of LoST is expected to have a positive impact on Swiss SMEs, particularly those in the fields of architecture, engineering, and product design. The technique enables more efficient and semantic coherent 3D shape tokenization, which can lead to significant cost savings and improved product design. Additionally, LoST can enable downstream tasks like semantic retrieval, which can be particularly useful for companies that rely on 3D models for product design and development. While the impact on the Swiss finance sector may be indirect, the development of LoST can lead to increased innovation and competitiveness in the Swiss economy.

Section 4 – What to Watch

The development of LoST is expected to be closely watched by the fintech and tech industries. CryptoForge plans to further develop and refine the technique, with the goal of making it more widely available. Investors and companies interested in 3D shape tokenization and AR modeling should keep a close eye on the company's progress and potential applications of LoST. Additionally, the development of LoST may lead to increased investment in the Swiss fintech sector, particularly in the areas of AI and machine learning.

Source

Original Article: LoST: Level of Semantics Tokenization for 3D Shapes

Published: March 18, 2026

Author: Niladri Shekhar Dutt

---

Disclaimer: This article is for informational purposes only and does not constitute financial advice. Consult a licensed financial advisor before making investment decisions.

Related Articles

• Toward Scalable Automated Repository-Level Datasets for Software Vulnerability Detection
• Finance AI in Finance Departments – How It Really Works
• Swiss SMEs & Accounting 2025: My Daily Experience with AI Tools