Sophistication in GenAI Use: Field Evidence from a Large Firm
研究通过分析一家大公司713,564条员工与生成式AI交互的数据，探讨了不同职能和层级员工使用生成式AI的复杂度差异及影响因素。
0 citationsRead paper
Selected work
Representative Papers
Quantum-enhanced satellite image classification
This work addresses the challenge of insufficient accuracy in multi-class satellite image classification for high-stakes remote sensing applications by proposing a hybrid quantum-classical approach. The method leverages many-body spin Hamiltonians to generate quantum features on an IBM quantum processor, which are then fused with a ResNet50 transfer learning model. To the best of our knowledge, this is the first demonstration of reproducible quantum-enhanced performance on a real-world satellite imagery task, overcoming recent limitations of near-term quantum devices in practical machine learning. Experimental results show that the proposed approach improves classification accuracy from the ResNet50 baseline of 83% to 87%, achieving an absolute gain of 2–3 percentage points.
0 citationsRead paper
Recent publications
Latest Papers
Sophistication in GenAI Use: Field Evidence from a Large Firm
研究通过分析一家大公司713,564条员工与生成式AI交互的数据，探讨了不同职能和层级员工使用生成式AI的复杂度差异及影响因素。
0 citationsRead paper
Quantum-enhanced satellite image classification
This work addresses the challenge of insufficient accuracy in multi-class satellite image classification for high-stakes remote sensing applications by proposing a hybrid quantum-classical approach. The method leverages many-body spin Hamiltonians to generate quantum features on an IBM quantum processor, which are then fused with a ResNet50 transfer learning model. To the best of our knowledge, this is the first demonstration of reproducible quantum-enhanced performance on a real-world satellite imagery task, overcoming recent limitations of near-term quantum devices in practical machine learning. Experimental results show that the proposed approach improves classification accuracy from the ResNet50 baseline of 83% to 87%, achieving an absolute gain of 2–3 percentage points.
0 citationsRead paper