As AI chips incorporate numerous parallelized cores to scale deep learning (DL) computing,
inter-core communication is enabled recently by employing high-bandwidth and low-latency …

To speed up computation, deep neural networks (DNNs) usually rely on highly optimized
tensor operators. Despite the effectiveness, tensor operators are often defined empirically …

Interest in intelligent edge computing is surging, driven by improving connectivity and
hardware advances. This is creating a need: today's cloud platforms optimize for high …

Real-world deep learning programs are often developed with dynamic programming
languages like Python, which usually have complex features, such as built-in functions and …

We present TensorBind, a novel approach aimed at unifying different hardware architectures
for compilation optimization. Our proposed framework establishes an embedding space to …

We are in the age of AI, with rapidly changing algorithms and hardware. As Deep Learning
(DL) applications have evolved, architects have turned to specialization to keep up with the …

Graph Neural Networks (GNNs) excel in tasks like node classification, edge predic-tion, and
graph classification, with applications in recommendation systems, network ana-lysis, and …

We present TensorBind, a novel approach aimed at unifying different hardware architectures
for compilation optimization. Our proposed framework establishes an embedding space to …