KGRKJGETMRETU895U-589TY5MIGM5JGB5SDFESFREWTGR54TY
Server : Apache/2.4.62
System : FreeBSD fbsdweb2.web.rcn.net 14.1-RELEASE FreeBSD 14.1-RELEASE releng/14.1-n267679-10e31f0946d8 GENERIC amd64
User : www ( 80)
PHP Version : 8.3.8
Disable Function : NONE
Directory :  /domains/highlandlabs/cqi-bin/ALFA_DATA/alfasymlink/root/domains/howardbender2/alci/ra/

Upload File :
current_dir [ Writeable ] document_root [ Writeable ]

 

Current File : /domains/highlandlabs/cqi-bin/ALFA_DATA/alfasymlink/root/domains/howardbender2/alci/ra/ra-1.doc
��ࡱ�>��	SU����R�������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������M 	��U&bjbj�=�=	&2�W�W�!d������lRRRRRRRf���8��fc
@""(JJJJJJ�������$� ��
RJJJJJ
<RRJJ
<<<JRRJRJ�<J�<<<x
�RR�J������fP�����3
0c
�I<I�<ffRRRR�Multilingual Natural Language Understanding 
Using Relationship Analysis

Current approaches to natural language understanding involve statistical analyses to select meanings from a �world knowledge� database, to interpret the contextual meaning of messages.  However, Any Language Communications has developed a novel system that uses the innate relationships of the words in a sensible message to determine the contextual meanings of all the words, without using world knowledge.  The methodology is called �Relationship Analysis� and includes a class/category structure of language concepts, a weighted inheritance system, a number language word conversion, and a tailored genetic algorithm to select the best of the possible word meanings. 

Relationship Analysis is a powerful language independent method and has been tested using machine translations with English, French, and Arabic as source languages and English, French, German, Hindi, and Russian as target languages.  By including a speech recognition system as a front end, Relationship Analysis can contextually analyze voice communications and code them for processing by an automated system.  This capability has applications in voice communications with a personal computer, home robots, and foreign language learning.

A simplified form of Relationship Analysis can also do sophisticated text analyses, in which concepts in the text are recognized irrespective of the text language.  Such analyses have been demonstrated using English and Arabic texts, with applications that include concept searches, email routing, and language-independent programmed instruction.  In the opposite way, Relationship Analysis can contextually analyze documents for the topics they contain and save the information locations in a semantic metadata index.  This index can be used to harvest information desired by users from complex queries. 

Here�s what Relationship Analysis is and how it works.

The Natural Intelligence of Language

Natural Language Understanding (NLU) is the analysis of communication between a �speaker� and a �listener�, whether those individuals are communicating via literature, voice, or another medium.  The listener interprets the speaker�s intentions, picking the one meaning for each of the words/phrases that best matches the overall meaning of the message.  Since people do this with apparent ease, the approach to computerizing NLU has been to mimic the human communications environment.  That environment has been assumed to be based on the listener�s �world knowledge�, gleaned from a lifetime of experiences.  For more than 10 years a major research effort has been undertaken to collect, categorize, and store this massive amount of often contradictory world knowledge information. The best analyses seem to rely on statistical methods, and nearly all NLU research in recent years has been to find the most successful statistical approach.

But what if human communication doesn�t rely on the listener�s analysis of the speaker�s intentions at all?  What if the speaker, in the construction of the message, has (invisibly) inserted the actual single meaning of each word/phrase within the message?  Then the listener can extract this �natural intelligence� from the message and recognize the overall meaning from the collection of individual meanings.  In other words, for sensible sentences the listener doesn�t need world knowledge, and computers don�t need it either.

Consider the following sentence:

They met at the bank.

This sentence is ambiguous and therefore can�t be understood. For NLU (either human or machine) to be successful, the sentence must be further explained, such as:

They met at the bank to withdraw money.
They met at the bank where the fishing was best.
They met at the bank of spotlights.

As these examples illustrate, a message is ambiguous unless each of the words/phrases has a single distinct meaning.  That meaning, that natural intelligence, makes sense - if the speaker didn�t construct the message with distinct meanings for each word/phrase, the message would be ambiguous and neither human nor computer could understand it.  Any Language Communications has found a way for computers to extract this natural intelligence through Relationship Analysis.


Relationship Analysis

The basis for Relationship Analysis is our premise that the meaning of each word/phrase in a message can be determined from the possible meanings of the other words/phrases in the message.  The relationships between these various possible meanings are clarified by using a copyrighted multilevel class/category structure of language concepts we�ve developed, containing 16 classes (such as Living Things, Human Society, Behavior & Ethics, etc.), and 864 categories (such as Sleeping, Eating, Medicine, etc.).  The �major� language families (Chinese, Germanic, Indic, Japanese, Malayo-Polynesian, Romance, Semitic, and Slavic), comprising the first language of over 70% of the world�s people, have been inspected for consistency and completeness with these classes/categories. 

Analysis of the class/category relationships involves a weighted inheritance system for the words/phrases, a specially developed number language, and a type of tailored genetic algorithm. It�s clear that Relationship Analysis is a semantic analysis, but we�ve also found that including some language-specific syntax information (such as dependent/independent phrase parsing, part-of-speech clarification, possessive information, etc.) is necessary to produce accurate results, and syntax guidance distinctly improves system performance. We�ve also found that Relationship Analysis is language independent � messages in a variety of natural languages have been analyzed using the same software.  The system has been tested with English, French, and Arabic as source languages and with English, French, German, Hindi, and Russian as target languages.  Major Relationship Analysis components are:

Weighted Inheritance System. Word/phrase meanings are initially assigned a weight based on their common interpretation in the dictionary.  For example, �hot� meaning �extremely warm� has a higher initial weight than �hot� meaning �radioactive�. However, these weights are adjusted depending on relationships with other words in the message.  We�ve also found that, occasionally, language processing requires additional analysis to link category relationships beyond those found in the hierarchy. 

Number Language. Words have always been difficult for computers to evaluate.  Consequently, each word/phrase entered in the system is transformed into a number that represents its relative place in the class/category organization.  By forming pairs of the words/phrases and comparing their relationship values, values for all message elements are obtained.  Such valuations are calculated for all the possible word/phrase meanings in the message.

Genetic Algorithm. The possible meanings for words quickly produce a massive number of possible message interpretations.  Even a seven-word sentence can easily result in over 100,000 possible sentence interpretations.  While only a few of the sentences will be �sensible�, the computer has no way of knowing which are and which aren�t, and the combinatorial explosion of the analysis can overrun the processing capabilities of most computers.  In fact, this was one of the major reasons for failure in early NLU attempts.  Recently, a mathematical technique called �genetic algorithms� was developed to address these �hard� problems, and has been applied to weather forecasting, pipeline analysis, traveling salesman problems, etc.  Conceptually similar to the way body cells produce DNA, the most viable products survive to combine with other viable products to produce the �fittest� final product.  In Relationship Analysis, partial message solutions are compared with each other, with the �best� ones remaining while the others are cast off.  Through multiple combinations and adjustments, the best message is developed.  This may be the first use of genetic algorithm methods for natural language analysis.

Summary

Relationship Analysis is a powerful semantics-oriented analysis technique that has produced contextually correct interpretations of words and phrases, and linguistically accurate machine translations of those words, in a variety of natural languages.  We have also found that Relationship Analysis can recognize nuances in messages not possible in purely syntactic approaches.  Identified potential applications for Relationship Analysis are natural language understanding, machine translation, text analysis, and semantic metadata indexing.



� 2004 Any Language Communications    www.anylanguage.net   Page 	 PAGE 4 of  NUMPAGES 4



HIJef�����ZwL]�#�#�#�%�%�%�%5&6&<&=&>&?&C&D&N&O&P&Q&U&������������������������º����º�hmHnHujUhh	jU
5�CJ\�CJ5�\�PJCJCJOJQJ^JOJQJ^J
6�CJ]�OJQJ^J$-IJ��		ef����qr���������������������������$
���!�1��o�]�1�^�o�a$$�1��o�]�1�^�o�a$$�1��o�]�1�^�o�a$	�1��o�]�1�^�o�

���1��o�]�1�^�o��%T&���ab�����������YZKL�#�#�#�#�%�%�������������������������$a$$�1��o�]�1�^�o�a$$
���!�1��o�]�1�^�o�a$�%�%�%R&S&T&U&������$a$

�H�1��o�]�1�^�o�$a$$&P1����/ ��=!��
"��	#��$��%��<Dd��gf�p�
�
C�LA�4�..\ALCimages\ALC-Logo.GIF��b��;��.��۬J��~�<2��;D�n�;��.��۬J��~�<2��PNG


IHDR�b�|1�PLTE���������������������������3f���3333f3�3�3�ff3fff�f�f���3�f��������3�f̙�����3�f������3333f3�3�3�3333333f33�33�33�3f3f33ff3f�3f�3f�3�3�33�f3��3��3��3�3�33�f3̙3��3�3�3�33�f3��3��3��ff3fff�f�f�f3f33f3ff3�f3�f3�ffff3fffff�ff�ff�f�f�3f�ff��f��f��f�f�3f�ff̙f��f�f�f�3f�ff��f��f����3�f���̙��3�33�3f�3��3̙3��f�f3�ff�f��f̙f�����3��f�����̙������3��f�̙��̙�����3��f�����̙����3�f�������3�33�3f�3��3��3��f�f3�ff�f��f��f�̙̙3̙f̙�̙�̙�����3��f�̙�������3�f��������3�f������3�33�3f�3��3�3��f�f3�ff�f��f�f�����3��f������������3��f�̙��������3��f��������Db�PbKGD�HcmPPJCmp0712Hs�8IDATx^�}˖�H�e~�@��+`�^+�����B� �	j���	�1	����{��H<��ɮ�Z��<�e�no777�r�YWv>���|���4�+8����/w[����ml�n
.z�j�������@���_}��*�0N�(~������b�˜�.��헠�h7�x�<�L
�ި�
�����%k�����p��c9߼ǟ~�g���h
9Y�LJ_��n�_1d����Qw���쮲ʒٰ�����μa\n��{��`��"�^_-B��Q�����g6}���ƽ^o���Y�ś���t���E�K����1P��n�����q�+�Qo���_t?]O#t�k�Q�1`E��X�eS���װ�@p�*,t���h�D�ao���M��Q����R��Y���^_h�>��k���ao�s��^�LHw	�Z�4�-\��$k��ǃ�zt��'���io4u���Ѱ�u*����]�I��W�n���;�^�A/�T��.�Qj��1�pAm��8+�XnS�9�M_��棏���<��G��]�Fإq�j���{fa������@~hJ��My�P�[��A���M֩��O}W�7g�b��
z����oT�w@?
%�{�Z������2o���Z�Ѯ��mA_!���Z5>Iڹ��ݶ�_�Ev���.��(�x��&?{Եn�5���6���=��"+2���.��u�@?H
ov�����m�Z����/��%t����A?��K�	lP��t�:MFc�3[���̵�����l�
%�oB:��+�^
}�(����
�q�T`���}��΅��~X���X{��UM������5J��w��9���c}tgdx��Vͱr�'E^��C-z��Y����rK�0;3Z������pm�?�}���m��3���`4�D�����+	9X
z<���7v�R��ݺ�Ѥ�����O5���@_W�צLщI�(�N������ � ?'T�!��f�n`@�
�U�����u��ɀ\�5��	��m�;�|ݧߟ����������)��q��{��A�v|��l˜�޸K�Z�k�Q�@�����u�o��/�+��x^�����+��зӼ�ܔ���E�ǜ�#%٣~‡B?����]�1t�KZ�A����?���H�o�ߎ؞���m��Q�7U�Q���-`ppM�����
ڝ�?�s���o-�����8ݟc
��Bd��E��qU=���_������PB�D��|���@�³�4�^U�Ո�{���̷դR<��m5�p1��ǬI�ms�o��9b���%��*��|T��r���#7?E`�6R9�Y�u��>��Q������?���W��~������:g�i�$��vI�e7M�����q��^/ԱW}q'>��؍R�"?oV��E~��$����>�S8t�����l�{Iv>%�8�e�0�4��PKfp��om,|0;����6�`��]��:Br�(�h���+t�'X�M�����c���������t��oL�|�E�n��^M)@
<4p1����e]I��8�`�S�Nod֦o#X�a\��ku��.N�GK���k�1�ɶ�G��S�
g)���1�E/"�h�zz�{r���&S�#��t���C���q�Ġ%p@��O���;Ĉz�=���&��XJ|m�-
�P�;l�D��r��}����[������Y.�G����Ο@!��[g��<��K_Z�8�zB��$	5ܿ��b��,���-���p�_ԛc�7JmWሑ���,&��N>�rH�1�}�a
��Qq���x,Ž!�7�6��g�w<��M���O0�!�"��"�#��'v�q�W��%����ђ�X�GKL4K6����$_z5?Z�D����m33����~
��wu�y:�}����=��C�q
��/��.��i��kEu=�e�@k�fPr��ި�62(
!�����4?�������� v4�Q1���wLs�O�G��;��R�kPv�
������O�i��Xۦ��y�EH�o�v7V�sLk�����p��?z�f�)������t���-���g��Q�� ��y��V&�s���fA��A���K>z�pTO�Ő���D���+v�q�"_���f=�D��I��6��r��Yx�A,��^��.;�'xd�;�Q���SGL��٠��d�3z��3�m�b�<>��[�}��uA��l�|1;�Y����y�9�υb[��q���L�s��p�e?���I�"'��٬7��<��7c��\�uq?;އ�1}�ȧ�	Ez3��^�!��?�ڦ�-μ�c�.����5�O�.�-F�O}:�z�(��],���h��
lٰ��Z"W3?�_�_%�3T��IE	��~�����_W,���Yo0���x:j��1�����5�p�s.hć�O���g���@1t�����s:�_�Sd�/5}�Μ�q2�z��Y1��Q�ꍝ�M$����R�(L��{�1((RmȔ\�Hef9}"@}��	*D=�{zra�TЈx&�z���t���u�ܸyM���
�V4�[�t������}*��o3��@\'�b&��	T�IX3b�2f�@�������Iӭ��^
�
5���͹�U5���^�C���9}(`�H���Y�Gk<��8$�p�+�ԃ�t:�����w'��ގ��cR� ř��#�����Z����=�Q��Ԫ��_\%�JJ��4�*�����p��L�<�{�y~�����3�K2��e��6I���a˲7�~���K,"����qf�	�C�G�f�z�� �}>e���G��ؗ�E�8+���-=��7�F0;�HAp	{�O�aOKPtr�Ap�+?���bwdBы����L[�-���S��,ay1x��O'b��h���k9���ь{3��2�}g��x^.~_����`I.��u��p�G�'�H+�ikwWO@��h\�j�u}-��i3�C�K�*D_�Y��5��4��	�3�����6��q�A�&e�`^yAa壯1`��&ָd=���}<H�?0l*E�F�qv�Z�B��$�__���5cӍ�N�w[@n�����5B�8E��y�����y��YP��)Wͽ�
G�*�p��"��b�7r[C�x� ���-�U������h9*��0	$��	4���!Dߡm�</g����![�F7<p**:������C��4���պf7�*�I�e��>���)�'-�Q��!����'��u����3L��,38v��������?��l���}a�0}'[\�2>�u<�	tt#t9/
���Z�^h�:g`��*#�l0��S�Z�����*+���O
.�Z��=q!�+M-���8��rP���ys`
�6u�~�8c:����cK]�Q��D�c���0�7�b�}�V�y��0�=�X2n�[�G#�����/���7΂�-R�*d	�So�'l�х���+gT��
��s�@/�K��*�=Q��s@�m�G��V!r�D`�-�-�c|:pT���fq��U8�=�r®"�)7A@s�V������rP�ϪY7�Q��MԲ��dl�ϪWh�~4E����1}�>{�
n���v��jf�Kn\�ݜ}$5~�2�k{)�7a�>�7g@s���]���ԟ�.�k{bw�t#Ѵ��KK��u�T
O/0������'n�tA�c+		�g���RoA��GJ���
KDH]%ߥpׄ�hx�!�f�te�
Z�F#(��P�/�%��h�_Fm�>6&���&IN��b8���MX�Q��N;
�{\�Fg����A:`���i��]��&����
R�vk���U�FkP�8&�`߻O�s�_��S7�_�a�t�FE��`���dB�>ɕ"�����E�Q9�Ƀu�绪���w0��IĞ�'6> �P��q7�ؑ�n1������I�ӵ���.�8�yA�ۑ�����<-:F�F��6��W=��Z

Asq(
����34��sqcb�4�C��蔴�	:���YN~+�r���{p����QBm�GN�º7VwT/�^��;}3+���AO~6u\I�>�]�ױW}v��$�>C��f�,h��3g��x%b@��)��ʵ�˸;�;��0��&p��R�N���'�|�����e���aty�]^��Ց@�o�\�/
�Ć������1�6B�!`:Q�ƒ�k��9&�G��d�WG|��#b8j�b{�@����g�N�z�gn�>(���}a������I�lv7u��}Z�3f��b�ƚ�K�j���
��TK�<�iܸ�ՄZ��J��-Ȝ[3�0�X[꿑<=�D3�[u^�C�bAټ#H�[�<?~��c�_��Zso�):�}�V�AȾޱ�>2uS,�Z��ʠ����l�d�0���?R*	p�@ſ�?4����7Ud���7�_�w�UO=?��~[�[��h^���W%�	�FΝ�Q���Wl�S�c/�i�!�jM�H,L��/.h&n��@���X������F�7��6�0��Z�.�gl>�n[L�AV���e��r��%�x���6l�o�
�`�q�ZM�d��{�Cٽ:
�3�6,�>�gKSm�Ij<BY�;-u��[���e��=�����6��0hLn���G��'Qu2_s�J=�-~f��7��{�OjD3k�d��gP�C�C�u�N���U8~��Կv�[�r…49����F�廹��G�,+R��߇h�a`|�"���E�<�;!�ρE�c츻�X�uj��+�r��{�@,n����}���Ĥ�H�D��`~��(Z �����˱��������/7���Jv�>�A�	�廃���M&����jŽ�����S����N6���X��y��O�E,"]�����^!�d:���e�D�h��Ioz��&C�@�>���00�_�����h(�vd�]�:	\���o��)ὡz�s��f�ưp���
�3��BtTY��cR�K�ؖ���j"xZ
��͏�b����Oە�T+�<�#���~<5fq��plE�\As
VA��Y*~&-��ʋ�A_�PH�Uޒ&�l:��3�j���;�3]�N5�K�P����3M?��8v��: �M����"���� �zT��h���+(�Y~�n5�h��(�.��s<1���|�;�}4�و�b�G�W����s�j}��L��f�'m#rc��?��$v/�>�/@�&<��'�Y�#�6[���k��\�2ܺY
7��M��z)BfH���ӃK�P���Ήj�@��J�QcO8.��'��@���t2���/V�:^�-c.&(h�:";q�#G� Ϥ��+�-�gd}��@:bLS�:�"�$�"�.��1-�
��#�CWN�AJ��6B��� �>�
�v}O ^8F\ؼZ�|ҝ�����O�{-ڄ�������_���)�SL!M��`<���7{GA��
y�LꉰE�Zn,	#Hn����U�hm_�Ҽ�l��{��>�h��F�_]�~���>c��g֣�/>jaLa�!�,���֊��5�>n�o	�b�W�9.���zO8y���ʕ����i���#��C~xh�J$�)x�k˓��j�5�{�j��x�����S�����#�8,�H�
��|#�6��r��$H�hk��D���&k!��m�Vf���J-ܘ
�%)S���E�"i��H�@��h
o+(=yj�B��[C��]�ˬ `ˬ����:�uX��bC��V��W�f��:p������~2�(����𶇆�L����G$��+L1/�j�=���~��H���m�
�[��9s��:�L�j�j�Cݔ4�>��a��38-?������X�ʽ|�cq�eAaCH�U#�_b�NLb!�3M�c�d�����+�~
4�.�,�]v�C�Q_a%��x���Dq�T$>���f�dU��9p����ƥ��33.�qжK*���8�!GYA��~w�X(�����i�|���L�ic|���;
�<��c@��l��vq����(!ٵBR2�ב��s��<�eZ�
�Z41um�J���M�]۲X��;�f�6E���i��D`us���kC:�=SX�N�$(�)��ہh�	�(�FY[���54m��D��B��؞A��RE�J���E#�����F���G��dC�QН`�g�XK��)ޭ;j�@wy��Zaq�j��׀��҆>�:��������ҵ������kD ��V��k�Jk\`�ۚ���d��}�����X��3���EO�/��tHp"�/u�vQ�_��)/���M�\!u�3&�����?w��@�ob�g`L'XR+������:A��P�(��_�l3M�Ue�=��N��W���ȼ~�.�/}X��(�ᾎTL�S�1��;��fv�uR�S8%�)�N��>�l�b���>)nn(7�c�� ����I��g��Or�P<ϵi���-�ApU��|/8��4N:ć���|/P�*�\���5���/�LK0ٮ�秆�գ��c%���D��
�]��z�U��\��Q|R���Ӎv �4�i�.Q�*���.STjX�����N��(��i�w�2'�F4�g��i[�W�=B��J��K|���\'�(o[]�>�\cL�U�~���3�2��{1p�^ۢ�&��޴v�px���v8R����!еX���ƋÆ]p��2~���;;`��f�`�ʲ����n�f����!�(�H�qk�b���g�=I��0�I�E+OgiP���pzU�/;������[t��-�u����W�üx�^_3���u@��[p̊P3�_�侠V�[�+�N(�
g�@���E����W��@|r�NG�]x!�iSy(�N�V�sWV�c��r�k�:t�Q�������G�X��J4}d2wA7F]B�8�m��ꥠ�|��e	�1Hz8e��<ea
�f�	����=��Q��{��ĩ�1��$,�UL%��/�G5�����n[��t�iۨ^��]ЭG1d(�q�Җ-ò�4���{,>���A	�Z~@�_�1�wK�ܥQc���pX���u��B*�繻h!�v��N����3�_oi��=����
�
t����ߜ�k�_td�m+��Oע��T�����Kj0oE��>�]�N6����OX���2|���阼	���O��D��ڪ��軞��)A̚��RSR��t�b<��Fu�s�C�_�ʐ-�#����n~Z�vMp/��u"L�j��m���b��W�Q������Q����H�(���gnʇn6栞C�K�t��A���5�|�.aj�\o@�^������e�GS��I��4N�j����?�
5�Տ�ǐ����١:"ea�.J@�G�����L�|V�H�����~M�4]5��L�Ÿ�x�3��H�&��/C\�Yǵ����<�u��p- D�n���;�.����:Ҫ�%�u��M�-�R`��W�Lۑ4��D<�P�s9�2�qW���οz����6C�e���yS�/�o�⪤�0fM�8,`�v�_�"�dРR�����]*�j(�����7����iHm��z�;Q4��CU�JR��rLU5.m_j�����
�����m�X��i�^-���w��d���?*]&k[��;���+�7�3ӎ#|&KvJ�a��U����d��ܟʚ�H[���j�T�n�Zo�)c�R��m�/W��/4<2��ј��r-��r
Q>TsDz	�p<#g�UR�B��֫F�42�'���`0��A�0jFo�ab�0�5��bl�/k��jJ-U�:��6gU�)I^�7R�`8�Djԋb�e�Аx����;յ.k!��ɕ������Dm�}���N�d�I���c�I�1�7�W�G��5��#��:�G�Pb#���ru-J5�;r�ߏ�dMT�f�k�a|Q0��<�e�����<_�T'Ȫ6�9�tp�c��{��y����	
c�k�-������)�I���L�P���=	�@��	Cf�I��0M���O"���c2�4\�e��X�b�1ל8U���@e\�|F�!q�<Al,8��D95l��2��:2�cF�ę�]�xl���!m3��bE�0�7���ۇ$��QA����7Q&>B�*��!��	cRJ@ÿH�a^	9���B�L��t~���H?5��H����X��q��9�DC�Hk�z���l0T<g��+~��(Q�Z=��S���I9�P�P�J��HL
Z�E�<d� '�=�PM<���b80\�l��i��\+SK�Ӭ7|A�d?ƌ��
bE��(Eѹ�
O��A`���MNW��?@�2�^����)V��a���B_I:�矃�p���|Ũ��?k��ų��y�	�:��	�E22��9�Ze�阏�����+�=d9 ���
�H�8Q����0s�OU��R

G���I?��4V|�TE%��]��/�̑�V�o�_B+s�5]�A�<�=�7=�ӥd.�9N��5���ě^�U�01j��Z�<�
"��g�&8g�Cd,�`�q�,�3��g��
ӣ ��%R������C2G�7ƈ�s�VP��*iDS�+�Y(���/�t�������mS�T0��jS�iy��r�4g�C8���Mp��]e�����i��A:��U��
E
O�#�?R�[���-Odm[�+0@ߏ�E��p��V� ;�(m�*)��~����CN��&�'X���C�)[C��=�����*���^���Խ����I):*\�n���Pk�P�@�s��˳Ep���!`�H�Um�\����\L�.��!�����CӡU%���<���R�]@:��1��%I`	��8Y���qy�JV�hY|P�&k�)%0_�����k��E�V�(�w�R��g�RJ��yy��3�R�b�s��Y#1$���)I��]N�yor,h�^z(�),5�đ,.;K����Pџ��(�(ԈˋAt���N�t�2�@e�'c0+���%�A�������R�	i8b6��$��G���0�j��4YM]�G܄;Z�q )C��#RT���1���4����M���j�H�Y�"�z�0V���D<
���P�(�?Sk{��|�B�R��]����妲�5ݛ?i�M�J���x8����=`�)�8���S�~�S�����δ���	_:�!rǙ�R����툭�@I^:;,�f�VjL���b�,N+V�<�#8��%J&Q�Se�U�������ぜa�K8�}չ0�UZs��<��{e s��#׍�8W��>��*
C��2J|:�Բ��0��Z-��Sq�
���
9��xR,P��W��%΅;m%��E<�D�f�6���5�,9�Ķ�rR9�˔M��r�`�j�F��G#�W擤$7j�a���y�o�*���ӲZ�X	�Zaи�	gj6DN���`���Eo���ɂ�`	�r��f8zL�)9U�x�=*�D0
�:��>��ך+"%���	����<
���Q���h�
N��K�N߆#���B���%O��'�n�%��\	SJʺz����o��(툳�hJ�5��Q�F�gb	�
q>��71t#4c�t�G��g{/�
'@��|2/M@9a�x� �p�DPؕ.����,L'��$�a�D�C��h�[�[NS4a�j	Ў�02��"1xSF]N;�P
p��dJz�e:]�sVq�\���qx�`S�1z�3���*�R���u�ɬG+_C"�1�
�RJ�R�{+*�	;��b'=�1	Q7�L�a)�BC���?�b��.tqU���.?��'U9����=4�ie�h�`�6��0W^������oiAY��g'%0���s1�F�	;�Ӹ����ł�)ު7� �2ʉ������ZI�������ŐC�PF��/�ԗ�z(����Χ28�2E�Gh�Y�/2 p^���J���O�D���at4���9���1K��#<�m,ɯ��Ҙ4FM�7|w|�5`�Ybv|�P�C�)𡆛�'nB���,<�ph��Id�c�6j'4<,2jjcd�@��һsGޱ����M�9KD���t��+���/]��o#��D>b:�3������2?��z�R�$FI�*XFp�=bk�O�(�ua�C�[��d�y�P����
_�a��YsRO�d�..��%��#���w���I͠�Q��OU�Ac�{	AO�Ǡ�JF��ȓ�Y=��+e��XH���>@��dE�i@�8�W�D�ӫo5W���z�K,�O��)v(�3,�hр�0BCS&83�p²�9�����ڷ=���b�C�U�҂^����.u�9 4��X%Cz-�@�TDU�h�c��G�5����ی�z�q\��S-|������Դ}0
c��8j均|<8�(	>�7�Ja c[�D��ǡ�cz�v�V����ox�z���4�
�0@��&�ޢ��p���'��9w8�_6�vɦ�3uT�C��%���q(���}���b�(�SV;�u�S��
,D��>#>S*��p��2h����r3��Qr�^f��X�A�i�8�C�K+ѽϯj�hgH��L�mW�מ ��d�#����6�`�a_�R��IcI	>��r)�1�T��E2�&���� �lX��}����A�obp�&��|�+��I���G�iS�
S�(s��ʝ�&�脋WrV'�q�-'QK@��x�'#q�fI��~�$�̮M3�F���	�hyp�Ɓ�‘|-f1 �V�0.�7���}E"(*��������-�#��a���Eɱ�G��"lqEKF=,�1\��������B��fR�/�~XڑD��<��$�}@ŗ�x�vY�%�~@兠6��X';�,]�ۯ�;����ͳ���@����>A���Ʊ�*J�.jY
�Wz�&u��%�����\6nK��������^hmWs<��[��?�?�N��wI�ˁUJv��cp��.-W�ӠK�@Z�����
CSE-��*����"��0w,�ƣ� &�8�.�%���\eg��ݨc�}��=b�9��T�2Vw��y�=�����3q�(y�w{#�$K�����'����B�^���j�b��r
��<

*�l,���r�`K�;��"�t[�a#6洨Qwx2�N�)�,?�(sx0�PP��\��ނ̌���Y谣�`ԁ*���#]��t���۴�1�����8�x	Y�}Y-���߹XG��&�n��"JtQG��%:�ֻ(�$�5s�S�����x�}�q!j�x`�A��58=d�b`��i
���D�+`��_p&f��K!pC	bo��B2�˼z�O0K`U)@5��e�>ګ`�B�1���&�WU+5_h�����G��IFc���Q��
zm��6��X�K��qT5�bV��t�;8���M�EJfmz�t���":��]8MB���#�3~���՟V$�.<�v�-&�B�������0,U��:�S���G�6@���_}�ϐTK���_��nSZ��0�����3:��ߣ%_e���Ԍ8X� �^τD��%#ճ!G�X�@ �Lŷb�
���Ag�9�u��5�Ӻ��d.y*E�~X*%e�"P�5�Ҟ`~�k�)`T�U'�:y�T
i�Q?��ς@�M����F�?�Qd?*A��c�X��d�0+p9�������R��ʔ�m�@��Ij�K:�0�Q�r�4���O�����!�ƹX�GHA����W.a���a:
�0�Py�k:��+���.��|١��Yȅ)"�I�����\��j��\�r��1�fʑ��j4^�WF��bJ��+y��0@[Q��3N���rЀ(VF��+F+Ko�����'�,Nj�D�!�G	�U�O����j�Ȕ5�v�=Zjс��֠;��r�-�P�� {�.^qW�f���M�e��Rd�*�����p�=����0�d ?��@.'W���T�����a�	�؛,x^,�y��	�qm�u�a�S����5'���^�Aʊ��yD\r'MLd�z�:o�����/�J�q���k���űP�G|��p�Y
?�e|�<q4���5'j"ct�1�+6K�!��P���>�/�%0�*̂Un���E�Q�y��{���55�n�������f#���qy ��pb _��8,0m����]�!ޮ�l4<��i�cі�
G<$�
b�BX\��W�Wf�=M�166�ۤC��޶�!�ً�����U�^1�
)g����;-�o��}��ǔC��2յ��1z9j̄۳��vh:2�	�{�=�*�j�~b��;F=�&�^*�G�g�����?�O^GL��a�������5�/�	!�2��/�2F�J2�e��3�
Q���菆\��9�g���r
��!3
L\Er��PR������'g���Lz�+	@�����N{s_��:��N��2Q�g����W�UeK�@~�&�y�>,����5~AxK`�\| �$���SGr�K&ܵh���`a̢���yj�`�sD���Q�_�݄ȝaN�m+ø#E/�H`�c��S����{\�2Eu��Z�R%P�6�_C�b6r8�h��G9`��p�l�,|O��
V�z� [�
qte����D&
[Ы\�I6�'�Y�"&��R�BPR��[
�2���A��u{�ru�J���$i��	�t'e7�l����׮�u]^�gu����k������T��}�;����wr%�ƍ{(��/��۸:�!��<�X����6�e�2�=
+�¿�W���pL�]�z�/�#�nx&�2k���<������Z���b4{��;V5��o?C�l���&H�����~���im|�Y���p�܁E�8؀�Go,����-�p�*��c�+��#T�=C����G0�i��L{�	�����2J���f(x�գ-(|F&N�Æ�?a�W?w��5``�Y[:�.�h���M��U<����҇�xDk���JP�3X�n��zN-�a�h9��\-!X�j��	.v{���O��P�m@�.�>�J�
���^M�q�0<TGv��h�k�5ܕk�P�@h�8mZ����1�@چ�y�H��:��V\N��� ����'���_q�=؉�Fg���U��VȆ|��Y6�a�G���0�5�HSk0� �%��������Co<�C	<�&��s���Ji�ї&!*Ehl#-oB�ꤠ>���a`��K����@���4�1q߳���E��4�
��\^`V'���N�p�U�?�X��JBҺ���*�3F
�-߅�]��	]��n��Q��4QxH8B=�x��q��c�^��OJ[�_���u�h��Z�>�=7Qi���.|t�&_�����
i.a� ��*L �Q_��bu�SMG8�r��Q{��؀JJ٨|[V�jL��ծu
�A�[U1=4s�UN�`V�1liX&T�{)����ǚۘl��㵶�+8z�N�h|�T�]��WA� ��D�w��`|8�g�����]�ڶ�2P�g�M�0���w*�U-o�]�:�K�\��e���OI���"��×�ٸ�D~��*Ę��cͱ_�I5ICF��(7Յ��;b��r8���JԖ������k���VMh�1�
�`��8�����B�k������}ex�Q_A�_=� �wż�A]9���F��A�qӀ��ֺ;�AwP������f�IEND�B`�
i@@�@Normal CJOJQJ^J_HmH	sH	tH	<@<	Heading 2$$@&a$
5�CJ\�<A@�<Default Paragraph Font:D�:
List Continue

&F�x.>@.Title$a$
5�CJ\�,@,Header

���!, @",Footer

���!8B@28	Body Text$a$OJQJ^J2P@B2Body Text 2CJPJ.U@�Q.	Hyperlink>*B*ph�U"2����-IJ��ef����qr�
�
�
�
�
�
ab�����������YZKL�����!�!�!�!R"V"�0���0���0���0���0���0���0���0���0���0���0���0���0���0���0���0���0���0���0���0���0���0���0���0���0���0���0���0���0���0��0���0���0���0���0���0���0���0���0���0���0���0���0���0���0���0���@0���@0���@0��
0cccfU&��%U&T&ELNS^`f!�������!�!�!"/"3"4"5"?"C"Q"S"V"�	�	�	�
�
�
�O4���!�!�!"/"3"4"5"?"C"Q"S"V"3333�� -C:\Any Language Communications\Web\ra\ra1.doc ,C:\Any Language Communications\Web\ra\ra.doc ,C:\Any Language Communications\Web\ra\ra.doc ,C:\Any Language Communications\Web\ra\ra.doc ,C:\Any Language Communications\Web\ra\ra.doc ,C:\Any Language Communications\Web\ra\ra.doc�3#	[����������������^��`���)�8����8^�8`���)�������^��`���)�����^�`���()�p����p^�p`���()��	�����	^��	`���()�@����@^�@`���.�������^��`���.�����^�`���.�3#	���������!Q"V"��@��ȉU"P@��Unknown������������G��z ��Times New Roman5��Symbol3&��z ��Arial?5�	�z ��Courier NewI&����������?�?Arial Unicode MS"1���h�*�s��f	��;!��
�Kf��0d\"2�q����,Multilingual Natural Language Understanding   ����Oh��+'��0�������	 ,
HT
`lt|��-Multilingual Natural Language Understanding 9.0ult ltltltNormal.dotl rm9rmMicrosoft Word 9.0l@`4<@��X]��@�~���������՜.��+,��D��՜.��+,��Xhp|�������
��� C;\"�	-Multilingual Natural Language Understanding Title� 8@_PID_HLINKS�Al	�%..\ALCimages\ALC-Logo.GIF	

���� !"#$%&'()*+,-./012345678����:;<=>?@A����CDEFGHI����KLMNOPQ��������T��������������������������������������������������������������������������������������������������������������������������������������������������������������������������������Root Entry��������	�FP����V�Data
�������������<1Table��������9IWordDocument����&2SummaryInformation(������������BDocumentSummaryInformation8��������JCompObj����jObjectPool������������P����P����������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������
����	�FMicrosoft Word Document
MSWordDocWord.Document.8�9�q

Anon7 - 2021