This document outlines a study presented at the 21st annual meeting of the Association for Natural Language Processing, focusing on a new approach to measure the similarity among language grammars to estimate the relativity of language learnability. The analysis includes 15 languages encoded against 24 grammatical features, highlighting the complexities and biases within the sampling. It aims to identify which language is most similar to Japanese, utilizing formal concept analysis for effective comparisons.
Astana, Kazakhstan Innovation Forum November 2012Ilkka Kakko
?
The document discusses the emergence of third generation science parks (3GSP) and innovative networks. It outlines trends driving changes in innovation landscapes that challenge existing science park models. 3GSPs focus on supporting communities and social capital to facilitate collaboration beyond institutions. They emphasize diversity, serendipity, and global connections through virtual platforms to attract talent and create new knowledge combinations. 3GSPs aim to cultivate "sticky knowledge" within professional communities to maintain regional competitiveness in hyper-turbulent environments.
Harnessing serendipity introduction into topic in miniworkshop, Coworking Eur...Ilkka Kakko
?
This document discusses harnessing serendipity in coworking spaces. It provides background on the concept of serendipity, tracing it to a Persian fairy tale. Examples of serendipitous discoveries in history are given from various domains. Different definitions of serendipity are examined, focusing on the elements of an unexpected encounter leading to an insight that provides value. The document explores ways to plan for and design opportunities for serendipity, such as through unexpected physical or virtual environments and encounters that support a serendipitous lifestyle and emergence in complex systems.
Kow Kuroda's talk at Concept Workshop, Japan Psychological Association 2011Kow Kuroda
?
The document discusses the nature and purpose of concepts, emphasizing their role in organizing perception and memory. It critiques the lack of operational definitions in existing literature and proposes a theoretical framework where concepts function as indices for managing vast amounts of memory. The author explores the implications of exceptional memory conditions and the balance between the benefits and burdens of memory recall.
“How to Support and Develop the Innovation-oriented Entrepreneurship in Turbu...Ilkka Kakko
?
The article discusses the 'vucability' approach to innovation management in response to the complexities of the postnormal VUCA era, highlighting the need for new strategies as traditional models are insufficient. It emphasizes three key dimensions of innovation management: serendipity thinking, platform utilization, and innovation ecosystems. The authors advocate for systemic evaluations in science and technology parks to adapt to the challenges posed by volatile, uncertain, complex, and ambiguous conditions.
"Platform thinking within the Third Generation Science Park Concept"; UNESCO ...Ilkka Kakko
?
The paper discusses the emergence of the Third Generation Science Park (3GSP) concept in Finland, emphasizing platform thinking as a critical factor in ecosystem development. It highlights the transformation of traditional innovation models, particularly through case studies like the Espoo Innovation Garden, which represents a shift towards more dynamic, co-created environments. The document outlines the essential features of competence platforms and their role in advancing innovation through collaboration, adaptability, and serendipity management.
A look inside the distributionally similar termsKow Kuroda
?
The document discusses distributional similarity and its relationship to semantic similarity. It presents three main questions about this relationship: 1) Can distributional similarity be equated with semantic similarity? 2) To what extent can they be equated? 3) Is the equating valid on a large scale? The document outlines the authors' study which involves selecting base terms, identifying their most distributionally similar terms, having human raters classify these term pairs into semantic relation categories, and analyzing the results.
Kow Kuroda's talk at JCLA 16 Workshop: Methodology in Cognitive Linguistics, ...Kow Kuroda
?
This advocates, in Japanese, Evidence-Based Linguistics (EBL) inspired by Evidence-Based Medicine (EBM) after discussing (unexpected) similarities between medical research and humanities. The main point is that if theoretical linguistics wants to be a truly empirical science, it definitely needs to rid itself from authority-based and overly intuition-based attitudes; otherwise, it remains to be a pseudo-science forever.
Kuroda NLP17 slides on Parallel Simulated Error Correction methodKow Kuroda
?
This talk expounds in Japanese how Parallel Simulated Error Correction (PSEC) can be successfully applied to natural language data that is often incomplete and/or inconsistent, thereby providing a natural ground for Pattern Lattice Model of linguistic knowledge. One of the best features of PSEC is that it can deals with parody very effectively, which is hard to describe in other frameworks or conceptions of linguistic knowledge.
Situations as attractors for semantic interpretationKow Kuroda
?
The document discusses a previous corpus analysis of the Japanese verb "osou" and its English translations. It found that the usage of the verb, both literal and figurative, can be described by a hierarchy of roughly 15 semantic frames or situations at the most specific levels. Over half the uses were metaphorical, indicating that more fine-grained analysis is needed to fully capture meaning. The frames were also found to be coarser than what average people understand from examples, showing the need for more precise and "realistic" semantic specifications.
Patter lattice as a model of human's language processingKow Kuroda
?
The document discusses using superposition rather than substitution to model linguistic composition. Superposition involves overlaying linguistic patterns such that elements are either identical or in an instance-of relationship, allowing for inconsistent information. This approach has desirable properties, representing elements redundantly-free while allowing flexible host structures. The document provides an example of composing three patterns A, B, and C using superposition and illustrates the results.
A look inside the distributionally similar termsKow Kuroda
?
The document discusses distributional similarity and its relationship to semantic similarity. It presents three main questions about this relationship: 1) Can distributional similarity be equated with semantic similarity? 2) To what extent can they be equated? 3) Is the equating valid on a large scale? The document outlines the authors' study which involves selecting base terms, identifying their most distributionally similar terms, having human raters classify these term pairs into semantic relation categories, and analyzing the results.
Kow Kuroda's talk at JCLA 16 Workshop: Methodology in Cognitive Linguistics, ...Kow Kuroda
?
This advocates, in Japanese, Evidence-Based Linguistics (EBL) inspired by Evidence-Based Medicine (EBM) after discussing (unexpected) similarities between medical research and humanities. The main point is that if theoretical linguistics wants to be a truly empirical science, it definitely needs to rid itself from authority-based and overly intuition-based attitudes; otherwise, it remains to be a pseudo-science forever.
Kuroda NLP17 slides on Parallel Simulated Error Correction methodKow Kuroda
?
This talk expounds in Japanese how Parallel Simulated Error Correction (PSEC) can be successfully applied to natural language data that is often incomplete and/or inconsistent, thereby providing a natural ground for Pattern Lattice Model of linguistic knowledge. One of the best features of PSEC is that it can deals with parody very effectively, which is hard to describe in other frameworks or conceptions of linguistic knowledge.
Situations as attractors for semantic interpretationKow Kuroda
?
The document discusses a previous corpus analysis of the Japanese verb "osou" and its English translations. It found that the usage of the verb, both literal and figurative, can be described by a hierarchy of roughly 15 semantic frames or situations at the most specific levels. Over half the uses were metaphorical, indicating that more fine-grained analysis is needed to fully capture meaning. The frames were also found to be coarser than what average people understand from examples, showing the need for more precise and "realistic" semantic specifications.
Patter lattice as a model of human's language processingKow Kuroda
?
The document discusses using superposition rather than substitution to model linguistic composition. Superposition involves overlaying linguistic patterns such that elements are either identical or in an instance-of relationship, allowing for inconsistent information. This approach has desirable properties, representing elements redundantly-free while allowing flexible host structures. The document provides an example of composing three patterns A, B, and C using superposition and illustrates the results.
20. 調査の枠組み ((現状))
20
FCAPreprocess
?
?
?
?
?
?
R.U.R.
in Czech
(Original)
R.U.R.
in English
R.U.R.
in French
R.U.R.
in
Japanese
Context of
Czech
word forms
Context of
English
word forms
Context of
Frech
word forms
Tokeninzed
version
Context of
Japanese
tokens
Concept Lattice
of Czech word
forms
Concept Lattice
of English
word forms
Concept Lattice
of Frech word
forms
Concept Lattice
of Japanese
tokens
Sunday, March 23, 14
21. 調査の枠組み ((展望))
21
FCAPreprocess
?
?
?
?
?
?
R.U.R.
in Czech
(Original)
R.U.R.
in English
R.U.R.
in French
R.U.R.
in
Japanese
Context of
Czech
word forms
Context of
English
word forms
Context of
Frech
word forms
Tokeninzed
version
Context of
Japanese
tokens
Concept Lattice
of Czech word
forms
Concept Lattice
of English
word forms
Concept Lattice
of Frech word
forms
Concept Lattice
of Japanese
tokens
Sunday, March 23, 14
24. 頻度の高い語末nn--ggrraammss
24
Rank
1 173 te 41 nou 16 osti 223 ng 216 ing 40 tion 21 ation
2 120 li 37 íte 16 ejte 172 ed 47 ers 35 ting 10 ction
3 117 ou 37 ?ch 14 nout 152 es 47 ion 21 ment 9 nding
4 112 ní 35 ali 13 ěj?í 127 er 45 ent 19 ning 8 ments
5 102 la 34 ého 13 ovat 84 ly 31 ess 19 hing 7 tions
6 86 ch 32 ají 13 ného 68 rs 27 est 17 ding 7 thing
7 82 em 28 ení 10 v?ch 64 st 26 red 17 ally 7 lling
8 74 ho 27 ili 10 vali 63 nt 25 ate 16 ying 7 ating
9 73 ně 25 ila 10 oval 61 on 25 lly 16 king 6 shing
10 72 at 25 vat 10 n?ch 61 ts 25 ted 15 ling 6 ities
11 70 it 24 eme 9 ství 56 le 24 ies 15 less 6 ility
12 68 me 23 uje 9 dají 54 al 23 ble 15 ight 6 ement
13 66 ím 22 ala 8 něte 54 re 22 nce 14 ring 6 aking
14 62 né 22 jte 8 luje 53 se 21 ght 13 able 5 tting
15 60 jí 20 ete 8 k?ch 51 ns 21 ine 12 ther 5 sting
16 59 lo 19 íme 7 ální 51 s 21 ter 12 sing 5 rying
17 58 ne 18 sti 7 ního 50 ce 19 ain 11 ters 5 right
18 56 ky 16 hle 7 jeme 48 en 19 ity 11 ions 5 owing
19 53 na 16 hne 7 jdou 47 ve 19 ive 11 ered 5 ology
20 51 ku 16 j?í 7 deme 45 te 18 der 11 ents 5 nting
21 49 ná 16 kou 7 dali 41 ne 18 one 10 ings 5 nging
22 48 le 16 out 6 áhne 40 ry 17 age 9 ture 5 ising
23 48 al 16 ího 6 títe 39 ss 17 ish 9 ness 5 inate
24 47 mi 15 –li 6 tane 37 ll 17 ver 9 ical 5 fully
25 43 ti 15 ním 6 ních 37 nd 16 her 9 ever 5 ening
26 42 ny 15 áte 6 nost 35 in 16 nts 9 ence 5 ching
27 42 ce 15 ěte 6 kého 35 ty 16 n’t 9 ated 5 cally
28 41 je 14 ují 6 kami 31 ge 16 ons 8 ving 5 aying
29 40 ?í 14 ích 6 jící 28 ar 15 ous 8 ssed
Eng 5gramsCzech 4 gramsCzech 2grams Czech 3grams Eng 2grams Eng 3grams Eng 4 grams
9 73 ně 25 ila 10 oval 61 on 25 lly 16 king 6 shing
10 72 at 25 vat 10 n?ch 61 ts 25 ted 15 ling 6 ities
11 70 it 24 eme 9 ství 56 le 24 ies 15 less 6 ility
12 68 me 23 uje 9 dají 54 al 23 ble 15 ight 6 ement
13 66 ím 22 ala 8 něte 54 re 22 nce 14 ring 6 aking
14 62 né 22 jte 8 luje 53 se 21 ght 13 able 5 tting
15 60 jí 20 ete 8 k?ch 51 ns 21 ine 12 ther 5 sting
16 59 lo 19 íme 7 ální 51 s 21 ter 12 sing 5 rying
17 58 ne 18 sti 7 ního 50 ce 19 ain 11 ters 5 right
18 56 ky 16 hle 7 jeme 48 en 19 ity 11 ions 5 owing
19 53 na 16 hne 7 jdou 47 ve 19 ive 11 ered 5 ology
20 51 ku 16 j?í 7 deme 45 te 18 der 11 ents 5 nting
21 49 ná 16 kou 7 dali 41 ne 18 one 10 ings 5 nging
22 48 le 16 out 6 áhne 40 ry 17 age 9 ture 5 ising
23 48 al 16 ího 6 títe 39 ss 17 ish 9 ness 5 inate
24 47 mi 15 –li 6 tane 37 ll 17 ver 9 ical 5 fully
25 43 ti 15 ním 6 ních 37 nd 16 her 9 ever 5 ening
26 42 ny 15 áte 6 nost 35 in 16 nts 9 ence 5 ching
27 42 ce 15 ěte 6 kého 35 ty 16 n’t 9 ated 5 cally
28 41 je 14 ují 6 kami 31 ge 16 ons 8 ving 5 aying
29 40 ?í 14 ích 6 jící 28 ar 15 ous 8 ssed
30 40 ek 13 ste 6 hází 27 ch 15 sed 8 ping
31 39 tí 13 tví 6 e?ně 26 ds 15 ves 8 lity
32 39 ci 13 ěla 26 ls 14 ned 8 ging
33 38 n? 12 alo 25 me 14 ure 8 ance
34 35 vá 12 ami 25 or 12 ded
35 35 st 12 dou 25 sh 12 ear
36 35 ka 12 kem 25 th 12 hes
37 33 nu 12 ost 24 an 12 ial
38 33 il 12 val 24 us 12 ide
39 32 ví 12 ání 12 ire
40 32 no 12 ngs
41 31 ni 12 ose
42 31 mu
Type freq 374 1798 3252 201 293 1839 2549
Sunday, March 23, 14