8 papers indexed
Afzal Ahmad, Gaoyu Mao, Shoubo Hu, Hui-Ling Zhen, Mingxuan Yuan, Xinyu Chen, et al.
Multi-task inference models share a single backbone across diverse tasks, yet execute identical computation regardless of which task is active - wasting energy and cycles on task-irrelevant operations. We observe that the task command, typically available before inference begins,…
Mengli Li, LinHan Fan, Yilei Fu, Yihe Yue, Xinyu Chen, Yi-Xiang Wang, et al.
Stretchable electrodes with strain-insensitive performance across arbitrary directions and geometries are essential for next-generation customizable wearable electronics. However, conventional designs based on ordered structures such as serpentines, kirigami, and island-bridge ge…
Yuhan Zhu, Changlian Ma, Xiangyu Zeng, Xinhao Li, Zhiqiu Zhang, Songze Li, et al.
Video multimodal large language models (MLLMs) can describe what happens in a video, but rarely identify when the supporting evidence occurs. We study generalist video temporal grounding, in which one model predicts a variable-cardinality set of evidence intervals across video le…
Xinhao Li, Yuhan Zhu, Xiangyu Zeng, Yuhao Dong, Haoning Wu, Zhiqiu Zhang, et al.
Recent advances in video understanding have spanned motion, long video, and streaming interaction, driving this field toward real-world applications. Despite this progress, current open-source models remain limited in several ways. They often struggle to generalize across diverse…
Financial observations are continuous, heterogeneous, and noisy, whereas decoder-only next-token models are usually built around discrete symbolic inputs. We introduce Vector-Input Autoregressive Inference for Ordinal-Return Modeling (VAIOM), a decoder-only Transformer for probab…
Jiawei Liang, Haotong Qin, Linfeng Du, Xingyu Liu, Shangkun Li, Hui Yu, et al.
Achieving nanosecond-scale inference latency for deep neural networks (DNNs) has become a primary architectural concern for latency-critical applications. While Field-Programmable Gate Arrays (FPGAs) offer a promising substrate for low-latency inference, conventional FPGA acceler…
Xinyu Chen, Yuqi Li, Jiabao Li, Pinyan Tang, Chong Wang, Aditi Majumder
Establishing correspondence between projector and camera images in a procam (projector + camera) system is essential for achieving high-resolution pixel matching, referred to as procam registration. The highest accuracy is typically obtained using structured light patterns (e.g.,…
Jiwen Ren, Yuming Xiong, Xinyu Chen, Yong Hao
The success of near-infrared spectroscopy (NIRS) analysis hinges on the precision and robustness of the calibration model. Shallow learning (SL) algorithms like partial least squares discriminant analysis (PLS-DA) often fall short in capturing the interrelationships between adjac…