Related papers: Integrated electro-optic attention nonlinearities …
Transformers with linear attention allow for efficient parallel training but can simultaneously be formulated as an RNN with 2D (matrix-valued) hidden states, thus enjoying linear-time inference complexity. However, linear attention…
Vision Transformers have achieved impressive performance in video classification, while suffering from the quadratic complexity caused by the Softmax attention mechanism. Some studies alleviate the computational costs by reducing the number…
Optical resonances in nanostructures enable strong enhancement of nonlinear processes at the nanoscale, such as second-harmonic generation (SHG), with high-$Q$ modes providing intensified light--matter interactions and sharp spectral…
The attention module is the key component in Transformers. While the global attention mechanism offers high expressiveness, its excessive computational cost restricts its applicability in various scenarios. In this paper, we propose a novel…
Thin-film lithium niobate (TFLN) has emerged as a promising platform for the realization of high performance chip-scale optical systems, spanning a range of applications from optical communications to microwave photonics. Such applications…
The ever-increasing demand for processing data with larger machine learning models requires more efficient hardware solutions due to limitations such as power dissipation and scalability. Optics is a promising contender for providing lower…
Investigating physical models with photonic synthetic dimensions has been generating great interest in vast fields of science. The rapid developing thin-film lithium niobate (TFLN) platform, for its numerous advantages including high…
Large transformer models have achieved state-of-the-art results in numerous natural language processing tasks. Among the pivotal components of the transformer architecture, the attention mechanism plays a crucial role in capturing token…
Thin-film lithium niobate (TFLN) enables extremely high-efficiency second-order nonlinear optical effects due to large nonlinear coefficient d33 and strong optical field localization. Here, we first designed and fabricated a…
The mid-infrared (MIR) spectral regime is central to applications including remote sensing, precision spectroscopy, higher harmonic generation, and free-space optical communication. However, coherent and broadband MIR modulation remains…
Achieving low optical loss is critical for scaling complex photonic systems. Thin-film lithium niobate (TFLN) offers strong electro-optic and nonlinear properties in a compact platform, making it ideal for quantum and nonlinear optics.…
Transformer-based language models utilize the attention mechanism for substantial performance improvements in almost all natural language processing (NLP) tasks. Similar attention structures are also extensively studied in several other…
Thin-film periodically poled lithium niobate (TF-PPLN) devices have recently gained prominence for efficient wavelength conversion processes in both classical and quantum applications. However, the patterning and poling of TF-PPLN devices…
Transformer-based deep learning models have achieved state-of-the-art performance across numerous language and vision tasks. While the self-attention mechanism, a core component of transformers, has proven capable of handling complex data…
Transformer architectures have achieved remarkable success in various domains. While efficient alternatives to Softmax Attention have been widely studied, the search for more expressive mechanisms grounded in theoretical insight-even at…
A novel thin-film LiNbO3 (TFLN) electro-optic modulator is proposed and demonstrated. LiNbO3-silica hybrid waveguide is adopted to maintain low optical loss for an electrode spacing as narrow as 3 {\mu}m, resulting in a record low…
Recently, the study of integrable Hamiltonian systems has led to nonlinear accelerator lattices with one or two transverse invariants and wide stable tune spreads. These lattices may drastically improve the performance of high-intensity…
The original softmax-based attention mechanism (regular attention) in the extremely successful Transformer architecture computes attention between $N$ tokens, each embedded in a $D$-dimensional head, with a time complexity of $O(N^2D)$.…
Advances in integrated photonics have enabled unprecedented level of control of light, powering a wide range of photonic technologies from communications and computing to precision metrology and quantum information. However, the…
The attention mechanism is an important reason for the success of transformers. It relies on computing pairwise relations between tokens. To reduce the high computational cost of standard quadratic attention, linear attention has been…