The science of flavour: why food researchers are interested in machine learning
Posted on May 19 2021 by Thomas Martin Walker
Story written by Dr Sarah Keenihan, AIML
Eveline van Honk, PhD Candidate
Freshly ground coffee. Ripe raspberries warmed in the sun. Crispy bacon. Aged red wine.
现磨咖啡。成熟的树莓在阳光下晒得暖烘烘的。酥脆的培根。陈年红酒。
What is it about some foods and drinks that make them so irresistible?
是什么让一些食物和饮料如此让人无法抗拒?
Fundamentally, it’s chemistry.
从根本上说，这是化学反应。
But there’s more to it than just that – dining is a complex human experience. Molecules in meals and beverages interact with receptors in our noses, mouths, on our tongues and in our upper airways and gut to create a response in our brains.
但还有比这更重要的——用餐是一种复杂的人类体验。食物和饮料中的分子与我们鼻子、嘴巴、舌头、上呼吸道和肠道中的受体相互作用，在我们的大脑中产生反应。
PhD candidate Eveline van Honk wants to apply artificial intelligence (AI) and machine learning to drill down into the science of flavour.
博士候选人伊芙琳·范·洪克希望应用人工智能(AI)和机器学习来深入研究风味科学。
She is part of a joint postgraduate program between the University of Adelaide and University of Nottingham UK, and will work with Professor Ian Fisk , Assistant Professor Ni Yang and Associate Professor Nicholas Watson in the UK, and experts in the School of Agriculture, Food and Wine (Associate Professor Sue Bastian) and the Australian Institute for Machine Learning (Dr Lingqiao Liu) in Australia.
她是阿德莱德大学和英国诺丁汉大学联合研究生项目的一部分，将与英国的伊恩·菲斯克教授、杨妮助理教授和尼古拉斯·沃森副教授，以及澳大利亚农业、食品和葡萄酒学院(副教授苏·巴斯蒂安)和澳大利亚机器学习研究所(刘凌桥博士)的专家合作。
“In food science, consumer experience is the most important factor, but using traditional approaches to measure this is laborious, time consuming and expensive,” van Honk says.
“在食品科学中，消费者体验是最重要的因素，但使用传统的方法来衡量这一点是费力的，耗时的和昂贵的，”范洪克说。
“Machine learning is a new approach that can help us get the most out of our data.”
“机器学习是一种新方法，可以帮助我们充分利用数据。”
Machine learning is a form of AI that works with large data sets to make predictions – in this case, what foods and drinks might taste like based on their chemical characteristics.
机器学习是人工智能的一种形式，它与大型数据集一起进行预测——在这种情况下，根据食物和饮料的化学特性，预测它们的味道。
“If machine learning can be applied to predict sensory attributes and consumer perception of different foods, then future product development and formulation could be accelerated, and resources saved,” van Honk explains.
“如果机器学习可以应用于预测感官属性和消费者对不同食品的感知，那么未来的产品开发和配方可以加速，并节省资源，”范·洪克解释说。
Van Honk has started her research in Nottingham, and hopes to travel to South Australia in the near future.
范洪克已经在诺丁汉开始了她的研究，并希望在不久的将来去南澳大利亚旅行。
Currently on secondment at the University of Adelaide, van Honk’s research co-supervisor Fisk runs the Food Flavour group at the University of Nottingham.
范洪克的研究合作主管菲斯克目前被阿德莱德大学借调，负责诺丁汉大学的食品风味小组。
“We’re interested in replicating the way that humans taste foods,” Fisk says.
菲斯克说:“我们感兴趣的是复制人类品尝食物的方式。”
“That experience is about smell, it’s about taste, it’s about texture – lots of different attributes come together to shape our experience of food and drink.”
“这种体验是关于气味、味道、质地的——许多不同的属性聚集在一起，塑造了我们对食物和饮料的体验。”
Fisk said if they can develop technology that pulls all of those experiences together, then in the future they may be able to predict what new food and drink products smell and taste like, and tailor food chemistry to match human preferences – for example in coffee or wine.
菲斯克说，如果他们能开发出将所有这些体验结合在一起的技术，那么在未来，他们可能能够预测新的食品和饮料产品的气味和味道，并根据人类的喜好定制食品化学——例如咖啡或葡萄酒。
“We think AI and machine learning are going to be vital in developing this capability,” says Fisk.
菲斯克说:“我们认为人工智能和机器学习对开发这种能力至关重要。”
For more information about jointly awarded PhD opportunities through the University of Adelaide, including with the University of Nottingham, please follow this link.
有关阿德莱德大学(包括诺丁汉大学)联合授予博士学位的更多信息，请点击此链接。
原文出处：
The science of flavour: why food researchers are interested in machine learning | Australian Institute for Machine Learning (AIML) | University of Adelaide
