Title: Diffusion Models for Increasing Accuracy in Olfaction Sensors and Datasets Show Chinese title

Title: 扩散模型在提高嗅觉传感器和数据集精度中的应用

Abstract: Robotic odour source localization (OSL) is a critical capability for autonomous systems operating in complex environments. However, current OSL methods often suffer from ambiguities, particularly when robots misattribute odours to incorrect objects due to limitations in olfactory datasets and sensor resolutions. To address this challenge, we introduce a novel machine learning method using diffusion-based molecular generation to enhance odour localization accuracy that can be used by itself or with automated olfactory dataset construction pipelines with vision-language models (VLMs) This generative process of our diffusion model expands the chemical space beyond the limitations of both current olfactory datasets and the training data of VLMs, enabling the identification of potential odourant molecules not previously documented. The generated molecules can then be more accurately validated using advanced olfactory sensors which emulate human olfactory recognition through electronic sensor arrays. By integrating visual analysis, language processing, and molecular generation, our framework enhances the ability of olfaction-vision models on robots to accurately associate odours with their correct sources, thereby improving navigation and decision-making in environments where olfactory cues are essential. Our methodology represents a foundational advancement in the field of robotic olfaction, offering a scalable solution to the challenges posed by limited olfactory data and sensor ambiguities. Show Chinese abstract

Abstract: 机器人气味源定位（OSL）是自主系统在复杂环境中的一项关键能力。然而，当前的OSL方法常常受到歧义的影响，特别是在由于嗅觉数据集和传感器分辨率的限制，机器人错误地将气味归因于不正确的物体时尤为明显。 为了解决这一挑战，我们引入了一种新颖的机器学习方法，该方法利用基于扩散的分子生成来提高气味定位的准确性，这种方法可以单独使用，也可以与自动化嗅觉数据集构建管道结合使用，其中包含视觉-语言模型（VLM）。我们的扩散模型的生成过程扩展了化学空间，超越了当前嗅觉数据集和VLM训练数据的局限性，从而能够识别出之前未记录的潜在气味分子。 然后，这些生成的分子可以通过先进的嗅觉传感器更准确地验证，这些传感器通过电子传感器阵列模拟人类嗅觉识别。 通过整合视觉分析、语言处理和分子生成，我们的框架增强了嗅觉-视觉模型在机器人上的能力，使其能够更准确地将气味与其正确来源关联起来，从而改善在气味线索至关重要的环境中的导航和决策。 我们的方法代表了机器人嗅觉领域的一项基础性进步，为解决由有限的嗅觉数据和传感器歧义所提出的挑战提供了可扩展的解决方案。