ASR

ASR knowledge transform

1. ASR technology practice.

2. Work for the next week

Voice interface based on speech is not possible right now, due to the limit of the Tencent SDK.
Voice API
1. Rich middle-layer architecture. Confirmed.
2. MLA functions: (1) single socket connection. simple test done. (2) connection distribution pooling (3) connection serialization (4) connection synchronization[Client selectable (syncrhonized/asynchronized), middle-layer synchronized].
3. MLA advanced functions: (1) connection authentication (2) connection investigation (heart beat)

Knowledge base search.
Expression formula representation, possibly using YACC to convert BNFs to trees.
a demo in 4 weeks.