Search for: All records

Tryptophan and its non‐canonical variants play critical roles in pharmaceutical molecules and enzymes. Facile access to this privileged class of amino acids from readily available building blocks remains a long‐standing challenge. Here, we report a regioselective synthesis of non‐canonical tryptophans bearing C4‐C7 substituents via Rh‐catalyzed annulation between structurally diverse tert‐butyloxycarbonyl (Boc)‐protected anilines and alkynyl chlorides readily prepared from amino acid building blocks. This transformation harnesses Boc‐directed C–H metalation and demetalation to afford a wide range of C2‐unsubstituted indole products in a redox‐neutral fashion. This umpolung approach compared to the classic Larock indole synthesis offers a novel mechanism for heteroarene annulation and will be useful for the synthesis of natural products and drug molecules containing non‐canonical tryptophan residues in a highly regioselective manner. 

Recently, deep neural networks (DNN) have been widely used in speaker recognition area. In order to achieve fast response time and high accuracy, the requirements for hardware resources increase rapidly. However, as the speaker recognition application is often implemented on mobile devices, it is necessary to maintain a low computational cost while keeping high accuracy in far-field condition. In this paper, we apply structural sparsification on time-delay neural networks (TDNN) to remove redundant structures and accelerate the execution. On our targeted hardware, our model can remove 60% of parameters and only slightly increasing equal error rate (EER) by 0.18% while our structural sparse model can achieve more than 1.5× speedup.