XJTLU projects on bird flight paths, AI music impress at CHI 2025

Innovative projects to untangle the complex flight paths of migratory birds and build an AI music creation system earned Xi¡¯an Jiaotong-Liverpool University honours at CHI 2025, a premier international conference on human-computer interaction.

43rd ACM CHI Conference on Human Factors in Computing Systems

An XJTLU team led by Dr Lingyun Yu from the School of Advanced Technology received the ¡°Best Paper¡± award for RouteFlow, a novel visualisation method for trajectories.

Taking inspiration from city transit management systems, RouteFlow uses algorithms to map the numerous overlapping flight paths of migratory birds and identify key stopping points.

Images 1 and 2 show the overlapping trajectories of bird migrations. Images 3 and 4 demonstrate the effect when RouteFlow is applied, making the trajectories and stopping points clearly visible.

The method is based on a hierarchical edge bundling algorithm, which clusters similar trajectories together to reveal overarching trends, and an incremental circular packing algorithm to organise hotspots on a map, preventing overlap and enhancing details.

Dr Lingyun Yu and students at CHI 2025

The conference also highlighted the work of a team led by Wanfang Xu, a master¡¯s student in XJTLU¡¯s MRes Computer Science programme, to develop an artificially intelligent music creation system that supports multimodal input, allowing users to express creative ideas through sounds, images, or real-time interactions.

Wanfang Xu presents her team¡¯s project at CHI 2025

¡°We face a common challenge when using AI to compose music: we can request a gentle melody but receive an intense rock piece instead,¡± says Xu. ¡°This happens because most AI music systems rely on text-based input, which doesn¡¯t always effectively convey emotions, imagery, or melodies.¡±

Using her team¡¯s system, users can hum a melody, play a cat¡¯s meow, or upload a seaside sunset photo, and the AI will translate it into a fitting piece of music. ¡°The system will also highlight the moment the input occurs, allowing users to track the sound,¡± she adds. ¡°By providing visual feedback, this helps users understand the AI¡¯s creative logic and make adjustments.¡±

Visualisation of music generated by inputting the sound of waves, an image of a cat, and text conveying ¡°happiness¡±.

Xu plans to extend this technology into other artistic areas to give users more intuitive control over their creations, transforming AI into a ¡°true creative partner¡±.

Both projects were the result of collaborations: RouteFlow was developed in partnership with Tsinghua University in China and Newcastle University in the United Kingdom; the AI music system originated from a workshop co-hosted by XJTLU, Duke Kunshan University, and the Beijing Institute of Technology.

CHI 2025, shorthand for the 43rd ACM CHI Conference on Human Factors in Computing Systems, was held in Yokohama, Japan, from 26 April to 1 May.

By Huatian Jin

Translated by Xueqi Wang

Edited by Patricia Pieterse

Photo courtesy of the School of Advanced Technology