ALBERT

All Library Books, journals and Electronic Records Telegrafenberg

feed icon rss

Your email was sent successfully. Check your inbox.

An error occurred while sending the email. Please try again.

Proceed reservation?

Export
  • 1
    Monograph available for loan
    Monograph available for loan
    Cambridge [u.a.] : MIT Press
    Associated volumes
    Call number: PIK M 031-94-0246
    In: Parallel distributed processing
    Type of Medium: Monograph available for loan
    Pages: XII, 611 S.
    Edition: 9. print.
    ISBN: 0262631105
    Location: A 18 - must be ordered
    Branch Library: PIK Library
    Location Call Number Expected Availability
    BibTip Others were also interested in ...
  • 2
    Monograph available for loan
    Monograph available for loan
    Cambridge [u.a.] : MIT Press
    Associated volumes
    Call number: PIK M 031-94-0245
    In: Parallel distributed processing
    Type of Medium: Monograph available for loan
    Pages: XX, 547 S.
    Edition: 9. print.
    ISBN: 026268053x
    Location: A 18 - must be ordered
    Branch Library: PIK Library
    Location Call Number Expected Availability
    BibTip Others were also interested in ...
  • 3
    ISSN: 0885-6125
    Keywords: Graded state machines ; finite state automata ; recurrent networks ; temporal contingencies ; prediction task
    Source: Springer Online Journal Archives 1860-2000
    Topics: Computer Science
    Notes: Abstract We explore a network architecture introduced by Elman (1990) for predicting successive elements of a sequence. The network uses the pattern of activation over a set of hidden units from time-step t-1, together with element t, to predict element t + 1. When the network is trained with strings from a particular finite-state grammar, it can learn to be a perfect finite-state recognizer for the grammar. When the net has a minimal number of hidden units, patterns on the hidden units come to correspond to the nodes of the grammar; however, this correspondence is not necessary for the network to act as a perfect finite-state recognizer. Next, we provide a detailed analysis of how the network acquires its internal representations. We show that the network progressively encodes more and more temporal context by means of a probability analysis. Finally, we explore the conditions under which the network can carry information about distant sequential contingencies across intervening elements to distant elements. Such information is maintained with relative ease if it is relevant at each intermediate step; it tends to be lost when intervening elements do not depend on it. At first glance this may suggest that such networks are not relevant to natural language, in which dependencies may span indefinite distances. However, embeddings in natural language are not completely independent of earlier information. The final simulation shows that long distance sequential contingencies can be encoded by the network even if only subtle statistical properties of embedded strings depend on the early information. The network encodes long-distance dependencies by shading internal representations that are responsible for processing common embeddings in otherwise different sequences. This ability to represent simultaneously similarities and differences between several sequences relies on the graded nature of representations used by the network, which contrast with the finite states of traditional automata. For this reason, the network and other similar architectures may be called Graded State Machines.
    Type of Medium: Electronic Resource
    Location Call Number Expected Availability
    BibTip Others were also interested in ...
  • 4
    ISSN: 0885-6125
    Keywords: Graded state machines ; finite state automata ; recurrent networks ; temporal contingencies ; prediction task
    Source: Springer Online Journal Archives 1860-2000
    Topics: Computer Science
    Notes: Abstract We explore a network architecture introduced by Elman (1990) for predicting successive elements of a sequence. The network uses the pattern of activation over a set of hidden units from time-step t-1, together with element t, to predict element t+1. When the network is trained with strings from a particular finite-state grammar, it can learn to be a perfect finite-state recognizer for the grammar. When the net has a minimal number of hidden units, patterns on the hidden units come to correspond to the nodes of the grammar, however, this correspondence is not necessary for the network to act as a perfect finite-state recognizer. Next, we provide a detailed analysis of how the network acquires its internal representations. We show that the network progressively encodes more and more temporal context by means of a probability analysis. Finally, we explore the conditions under which the network can carry information about distant sequential contingencies across intervening elements to distant elements. Such information is maintained with relative ease if it is relevant at each intermediate step, it tends to be lost when intervening elements do not depend on it. At first glance this may suggest that such networks are not relevant to natural language, in which dependencies may span indefinite distances. However, embed dings in natural language are not completely independent of earlier information. The final simulation shows that long distance sequential contingencies can be encoded by the network even if only subtle statistical properties of embedded strings depend on the early information. The network encodes long-distance dependencies by shading internal representations that are responsible for processing common embeddings in otherwise different sequences. This ability to represent simultaneously similarities and differences between several sequences relies on the graded nature of representations used by the network, which contrast with the finite states of traditional automata. For this reason, the network and other similar architectures may be called Graded State Machines.
    Type of Medium: Electronic Resource
    Location Call Number Expected Availability
    BibTip Others were also interested in ...
  • 5
    Electronic Resource
    Electronic Resource
    Oxford, UK : Blackwell Publishing Ltd
    Annals of the New York Academy of Sciences 444 (1985), S. 0 
    ISSN: 1749-6632
    Source: Blackwell Publishing Journal Backfiles 1879-2005
    Topics: Natural Sciences in General
    Type of Medium: Electronic Resource
    Location Call Number Expected Availability
    BibTip Others were also interested in ...
  • 6
    ISSN: 1573-515X
    Keywords: atmospheric nitrogen deposition ; fertilizers ; Land Margin Ecosystems Research ; nitrogen loading ; Waquoit Bay ; wastewater ; watersheds
    Source: Springer Online Journal Archives 1860-2000
    Topics: Chemistry and Pharmacology , Geosciences
    Notes: Abstract World-wide eutrophication of estuaries has made accurate estimation ofland-derived nitrogen loads an important priority. In this paper we verifypredictions of nitrogen loads made by the Waquoit Bay Nitrogen LoadingModel (NLM). NLM is appropriate for watersheds with mixes of forested,agricultural, and residential land uses, and underlain by coarseunconsolidated sediments. NLM tracks the fate of nitrogen inputs byatmospheric deposition, fertilizer use, and wastewater disposal, and assignslosses of nitrogen from each source as the nitrogen is transported throughthe land use mosaic on the watershed surface, then through the underlyingsoils, vadose zones, and aquifers. We verified predictions of nitrogen loads by NLM in two independent ways.First, we compared NLM predictions to measured nitrogen loads in differentsubestuaries in the Waquoit Bay estuarine system. Nitrogen loads predictedby NLM were statistically indistinguishable from field-measured nitrogenloading rates. The fit of model predictions to measurements remained goodacross the wide range of nitrogen loads, and across a broad range in size(10–10,000 ha) of land parcels. NLM predictions were most precise whenspecific parcels were larger than 200 ha, and within factors of 2 for smallerparcels. Second, we used NLM to predict the percentage of nitrogen loads toestuaries contributed by wastewater, and compared this prediction to theδ15N signature distinguishable from N derived fromatmospheric or fertilizer sources. The greater the contribution ofwastewater, the heavier the δ15N value in groundwater. Thesignificant linear relation between NLM predictions of percent wastewatercontributions and stable isotopic signature corroborated the conclusionthat model outputs provide a good match to empirical measurements. Thegood agreement obtained in both verification exercises suggests that NLMis an useful tool to address basic and applied questions about how land usepatterns alter the fate of nitrogen traversing land ecosystems, and thatNLM provides verified estimates of the land-derived nitrogen exports thattransform receiving aquatic ecosystems.
    Type of Medium: Electronic Resource
    Location Call Number Expected Availability
    BibTip Others were also interested in ...
  • 7
    Publication Date: 2019-05-17
    Description: An extensive body of empirical research has revealed remarkable regularities in the acquisition, organization, deployment, and neural representation of human semantic knowledge, thereby raising a fundamental conceptual question: What are the theoretical principles governing the ability of neural networks to acquire, organize, and deploy abstract knowledge by integrating across many individual experiences? We address this question by mathematically analyzing the nonlinear dynamics of learning in deep linear networks. We find exact solutions to this learning dynamics that yield a conceptual explanation for the prevalence of many disparate phenomena in semantic cognition, including the hierarchical differentiation of concepts through rapid developmental transitions, the ubiquity of semantic illusions between such transitions, the emergence of item typicality and category coherence as factors controlling the speed of semantic processing, changing patterns of inductive projection over development, and the conservation of semantic similarity in neural representations across species. Thus, surprisingly, our simple neural model qualitatively recapitulates many diverse regularities underlying semantic development, while providing analytic insight into how the statistical structure of an environment can interact with nonlinear deep-learning dynamics to give rise to these regularities.
    Print ISSN: 0027-8424
    Electronic ISSN: 1091-6490
    Topics: Biology , Medicine , Natural Sciences in General
    Location Call Number Expected Availability
    BibTip Others were also interested in ...
  • 8
    Publication Date: 2009-11-25
    Print ISSN: 0027-8424
    Electronic ISSN: 1091-6490
    Topics: Biology , Medicine , Natural Sciences in General
    Location Call Number Expected Availability
    BibTip Others were also interested in ...
  • 9
    Publication Date: 2011-06-11
    Print ISSN: 1559-2723
    Electronic ISSN: 1559-2731
    Topics: Geography
    Published by Springer
    Location Call Number Expected Availability
    BibTip Others were also interested in ...
  • 10
    Publication Date: 2019-05-06
    Description: Climate warming is expected to mobilize northern permafrost and peat organic carbon (PP-C), yet magnitudes and system specifics of even current releases are poorly constrained. While part of the PP-C will degrade at point of thaw to CO2 and CH4 to directly amplify global warming, another part will enter the fluvial network, potentially providing a window to observe large-scale PP-C remobilization patterns. Here, we employ a decade-long, high-temporal resolution record of 14C in dissolved and particulate organic carbon (DOC and POC, respectively) to deconvolute PP-C release in the large drainage basins of rivers across Siberia: Ob, Yenisey, Lena, and Kolyma. The 14C-constrained estimate of export specifically from PP-C corresponds to only 17 ± 8% of total fluvial organic carbon and serves as a benchmark for monitoring changes to fluvial PP-C remobilization in a warming Arctic. Whereas DOC was dominated by recent organic carbon and poorly traced PP-C (12 ± 8%), POC carried a much stronger signature of PP-C (63 ± 10%) and represents the best window to detect spatial and temporal dynamics of PP-C release. Distinct seasonal patterns suggest that while DOC primarily stems from gradual leaching of surface soils, POC reflects abrupt collapse of deeper deposits. Higher dissolved PP-C export by Ob and Yenisey aligns with discontinuous permafrost that facilitates leaching, whereas higher particulate PP-C export by Lena and Kolyma likely echoes the thermokarst-induced collapse of Pleistocene deposits. Quantitative 14C-based fingerprinting of fluvial organic carbon thus provides an opportunity to elucidate large-scale dynamics of PP-C remobilization in response to Arctic warming.
    Print ISSN: 0027-8424
    Electronic ISSN: 1091-6490
    Topics: Biology , Medicine , Natural Sciences in General
    Location Call Number Expected Availability
    BibTip Others were also interested in ...
Close ⊗
This website uses cookies and the analysis tool Matomo. More information can be found here...