= Jürgen Schmidhuber
{c}
<Jürgen Schmidhuber> is a <computer scientist> whose research includes learning architectures for <recurrent neural networks>. He coauthored the1997<long short-term memory> paper with <Sepp Hochreiter> and the2000adaptive-forgetting paper with <Felix A. Gers> and <Fred Cummins>. These are team contributions documented by the original bylines.<Jürgen Schmidhuber homepage>[Author identification and research links]{ref} <Long Short-Term Memory by Sepp Hochreiter and Jürgen Schmidhuber>[Original title page]{ref} <Learning to Forget: Continual Prediction with LSTM by Felix A. Gers, Jürgen Schmidhuber and Fred Cummins>[MIT Press full author byline]{ref}
The1997architecture separates a persistent internal state from learned control of access to that state. A <LSTM input gate> controls new input, while a <LSTM output gate> controls how stored information influences output. This arrangement was designed to reduce the <vanishing gradient problem> that makes distant events difficult to associate during learning.<Long Short-Term Memory by Sepp Hochreiter and Jürgen Schmidhuber>[Abstract; Sections3–4: constant error flow and multiplicative gates]{ref}
An illustrative task presents a marked value early in a long stream, then asks for it after many irrelevant inputs. A network needs both a way to retain the relevant information and a way to protect it from later interference. The paper tested artificial long-delay tasks; success on those experiments is evidence about the architecture and learning procedure under specified conditions, not a guarantee of unlimited or error-free memory.
\Image[https://people.idsia.ch/~juergen/juergen2017icub.jpg]
{title=Jürgen Schmidhuber, photograph linked from his author homepage}
{description=External photograph identified in the Photos list of <Jürgen Schmidhuber homepage>.}
Back to article page
