LSTM can solve hard long time lag problems

Sepp Hochreiter; Jürgen Schmidhuber

LSTM can solve hard long time lag problems

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

Standard recurrent nets cannot deal with long minimal time lags between relevant signals. Several recent NIPS papers propose alternative methods. We first show: problems used to promote various previous algorithms can be solved more quickly by random weight guessing than by the proposed algorithms. We then use LSTM, our own recent algorithm, to solve a hard problem that can neither be quickly solved by random search nor by any other recurrent net algorithm we are aware of.

Original language	English (US)
Title of host publication	Advances in Neural Information Processing Systems
Publisher	Neural information processing systems foundation
Pages	473-479
Number of pages	7
ISBN (Print)	0262100657
State	Published - Jan 1 1997
Externally published	Yes

Cite this

@inproceedings{1c9a977302914d7ab700cc27a9c42684,

title = "LSTM can solve hard long time lag problems",

abstract = "Standard recurrent nets cannot deal with long minimal time lags between relevant signals. Several recent NIPS papers propose alternative methods. We first show: problems used to promote various previous algorithms can be solved more quickly by random weight guessing than by the proposed algorithms. We then use LSTM, our own recent algorithm, to solve a hard problem that can neither be quickly solved by random search nor by any other recurrent net algorithm we are aware of.",

author = "Sepp Hochreiter and J{\"u}rgen Schmidhuber",

note = "Generated from Scopus record by KAUST IRTS on 2022-09-14",

year = "1997",

month = jan,

day = "1",

language = "English (US)",

isbn = "0262100657",

pages = "473--479",

booktitle = "Advances in Neural Information Processing Systems",

publisher = "Neural information processing systems foundation",

}

TY - GEN

T1 - LSTM can solve hard long time lag problems

AU - Hochreiter, Sepp

AU - Schmidhuber, Jürgen

N1 - Generated from Scopus record by KAUST IRTS on 2022-09-14

PY - 1997/1/1

Y1 - 1997/1/1

N2 - Standard recurrent nets cannot deal with long minimal time lags between relevant signals. Several recent NIPS papers propose alternative methods. We first show: problems used to promote various previous algorithms can be solved more quickly by random weight guessing than by the proposed algorithms. We then use LSTM, our own recent algorithm, to solve a hard problem that can neither be quickly solved by random search nor by any other recurrent net algorithm we are aware of.

AB - Standard recurrent nets cannot deal with long minimal time lags between relevant signals. Several recent NIPS papers propose alternative methods. We first show: problems used to promote various previous algorithms can be solved more quickly by random weight guessing than by the proposed algorithms. We then use LSTM, our own recent algorithm, to solve a hard problem that can neither be quickly solved by random search nor by any other recurrent net algorithm we are aware of.

UR - http://www.scopus.com/inward/record.url?scp=0000370416&partnerID=8YFLogxK

M3 - Conference contribution

SN - 0262100657

SP - 473

EP - 479

BT - Advances in Neural Information Processing Systems

PB - Neural information processing systems foundation

ER -

LSTM can solve hard long time lag problems

Abstract

Other files and links

Fingerprint

Cite this