ebook img

Simple, Efficient and Neural Algorithms for Sparse Coding PDF

112 Pages·2015·4.35 MB·English
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview Simple, Efficient and Neural Algorithms for Sparse Coding

Simple,  Efficient  and  Neural   Algorithms  for  Sparse  Coding   Ankur  Moitra  (MIT)   joint  work  with  Sanjeev  Arora,  Rong  Ge  and  Tengyu  Ma B.  A.  Olshausen,  D.  J.  Field.  “Eme        rgence  of  simple-­‐cell  recepNve   field  properNes  by  learning  a  sparse  code  for  natural  images”,   1996 B.  A.  Olshausen,  D.  J.  Field.  “Eme        rgence  of  simple-­‐cell  recepNve   field  properNes  by  learning  a  sparse  code  for  natural  images”,   1996   break  natural  images  into  patches:   (collecNon  of  vectors) B.  A.  Olshausen,  D.  J.  Field.  “Eme        rgence  of  simple-­‐cell  recepNve   field  properNes  by  learning  a  sparse  code  for  natural  images”,   1996   break  natural  images  into  patches:   sparse  coding   (collecNon  of  vectors) B.  A.  Olshausen,  D.  J.  Field.  “Eme        rgence  of  simple-­‐cell  recepNve   field  properNes  by  learning  a  sparse  code  for  natural  images”,   1996   break  natural  images  into  patches:   sparse  coding   Proper2es:  localized,     (collecNon  of  vectors)   bandpass  and  oriented B.  A.  Olshausen,  D.  J.  Field.  “Eme    rgence  of  simple-­‐cell  recepNve   field  properNes  by  learning  a  sparse  code  for  natural  images”,   1996   break  natural  images  into  patches:   (collecNon  of  vectors) B.  A.  Olshausen,  D.  J.  Field.  “Eme    rgence  of  simple-­‐cell  recepNve   field  properNes  by  learning  a  sparse  code  for  natural  images”,   1996   break  natural  images  into  patches:   singular  value   decomposi2on   (collecNon  of  vectors) B.  A.  Olshausen,  D.  J.  Field.  “Eme    rgence  of  simple-­‐cell  recepNve   field  properNes  by  learning  a  sparse  code  for  natural  images”,   1996   break  natural  images  into  patches:   singular  value   Noisy!   decomposi2on   Difficult  to     interpret!   (collecNon  of  vectors) OUTLINE       Are  there  efficient,  neural  algorithms  for  sparse   coding  with  provable  guarantees? OUTLINE       Are  there  efficient,  neural  algorithms  for  sparse   coding  with  provable  guarantees?   Part  I:  The  Olshausen-­‐Field  Update  Rule     Ÿ  A  Non-­‐convex  FormulaNon     Ÿ  Neural  ImplementaNon     Ÿ  A  GeneraNve  Model;  Prior  Work

Description:
Simple, Efficient and Neural. Algorithms for Sparse Coding. Ankur Moitra (MIT) joint work with Sanjeev Arora, Rong Ge and Tengyu Ma
See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.