新算法模仿果蝇嗅觉，实现长期气味记忆

冲绳科技大学的研究团队开发了一种名为Spi-Fly的新算法，旨在解决现有电子鼻技术的关键缺陷。1该算法灵感来自果蝇大脑的工作原理——果蝇仅凭约14万个神经元的大脑（体积比罂粟籽还小）就能在瞬间识别大量气味，并长期保留嗅觉记忆。1

Kevin Max和Yang Shen主导的这项研究已发表在《神经形态计算与工程》期刊上。1传统电子鼻存在价格昂贵、检测范围有限等问题，尤其是当设备学习识别新气味后，往往会遗忘已有的旧气味记忆。1新算法的开发正是为了克服这些局限，让电子鼻具备与果蝇相似的持久嗅觉学习能力。1

Researchers at the University of the Ryukyus have developed a novel algorithm called Spi-Fly that replicates the fruit fly's exceptional ability to recognize and retain memories of numerous odors almost instantaneously.1 The fruit fly brain, which contains approximately 140,000 neurons and is smaller than a poppy seed, accomplishes this olfactory feat with remarkable efficiency.1 Kevin Max and Yang Shen, the researchers behind this work, sought to overcome significant limitations in current electronic nose technology, which tends to be costly, has a narrow detection range, and struggles to retain memories of previously learned scents while acquiring new ones.1

The findings from Max and Shen's research have been published in the journal Neuromorphic Computing and Engineering.1 By understanding and implementing the biological mechanisms underlying the fruit fly's olfactory system, the Spi-Fly algorithm represents a promising advance in developing more capable electronic sensing systems.