Sharedwww / job / Short.dviOpen in CoCalc
����;� TeX output 1999.11.12:1119������ǂ����6�;��䃍�������������	�7F
C��q
cmbxti10�Wil�8�liam��sA.�Stein�\'R���Uese�ar�ch��sSummary��(�����d��k�ff���������2ō��|4V���a����N��qcmbx12�Researc��Zh��
Summary��'̑����d���N�G�cmbx12�1��(XIn��u�tro��=duction���#���d�X�Qcmr12�My��Wresearc��rh�program�re
ects�the�essen�tial�in�terpla�y�b�S�et�w�een�abstract�theory�and�explicit������dmac��rhine��zcomputation�during�the�latter�half�of�the�t�w�en�tieth�cen�tury;��it�sits�at�the�in�tersection�����dof��recen��rt�w�ork�of�B.�Mazur,�1�K.�Rib�S�et,�R.�T��Va��rylor,�and�A.�Wiles�on�Galois�represen��rtations�with�����dw��rork�@vof�J.�Cremona,�b�N.�Elkies,�and�J.-F.�Mestre�on�explicit�computations�in��rv�olving�@vmo�S�dular�����dab�S�elian��[v��X�arieties.���My�w��rork�on�the�Birc�h�and�Swinnerton-Dy�er�conjecture�for�mo�S�dular�ab�elian�����dv��X�arieties��Rand�searc��rh�for�new�examples�of�mo�S�dular�icosahedral�Galois�represen�tations�has�led�����dme��to�disco��rv�er��and�implemen��rt�algorithms�for�explicitly�computing�with�mo�S�dular�forms.������d�2��(XOb�zjectiv��u�es���#���d�The��main�outstanding�problem�in�m��ry�eld�is�the�conjecture�of�Birc�h�and�Swinnerton-Dy�er�����d(BSD�I�conjecture),�jwhic��rh�I�ties�together�the�constellation�of�arithmetic�in�v��X�arian�ts�of�an�elliptic�����dcurv��re.��There�Qis�still�no�general�class�of�elliptic�curv�es�for�whic�h�the�full�BSD�	conjecture�����dis�@1kno��rwn�to�hold.�	9zApproac�hes�to�the�BSD�?�conjecture�that�rely�on�congruences�b�S�et�w�een�����dmo�S�dular��forms�are�lik��rely�to�require�a�deep�er�understanding�of�the�analogous�conjecture�for�����dmo�S�dular��ab�elian�v��X�arieties,�whic��rh�are�higher�dimensional�analogues�of�elliptic�curv�es.�����XAs��a�rst�step,���I��lha��rv�e��obtained�theorems�that�mak��re�p�S�ossible�computation�of�some�of�the�����darithmetic��win��rv��X�arian�ts�of�mo�S�dular�ab�elian�v��X�arieties.�zMy�ob��jectiv��re�is�to�nd�w�a�ys�to�explicitly�����dcompute��all�of�the�arithmetic�in��rv��X�arian�ts.���Cremona��has�en��rumerated�these�in�v��X�arian�ts�for�the�����drst��.few�thousand�elliptic�curv��res,�Pand�I���am�w�orking�to�do�the�same�for�ab�S�elian�v��X�arieties.�����dIt���is�hop�S�ed�that�this�w��rork�will�con�tin�ue�to�yield�theoretical�results.�
��I�ƃam�also�writing�����dmo�S�dular�dforms�soft��rw�are�dthat�I�.hop�e�will�b�e�used�b��ry�man�y�mathematicians�and�ha�v�e�practical�����dapplications��in�the�dev��relopmen�t��of�elliptic�curv��re�cryptosystems.�����XMy��long-range�goal�is�to�giv��re�a�general�h�yp�S�othesis,��yv��X�alid�for�innitely�man�y�ab�S�elian�����dv��X�arieties,��under��^whic��rh�the�full�BSD���conjecture�holds.�
~My�approac�h�in�v�olv�es�com�bining�����dEuler��`system�tec��rhniques�of�K.�Kato�and�K.�Rubin�with�visibilit�y�and�congruence�ideas�of�����dMazur��and�Rib�S�et.������d�3��(XMo��=dular�z�ab�elian�v���arieties���#���d�My��:primary�ob��jectiv��re�is�to�v�erify�the�BSD��.conjecture�for�sp�S�ecic�mo�dular�ab�elian�v��X�arieties,�����db��ry��using�the�ric�h�theory�of�their�arithmetic.�����XThe�M�BSD�M�conjecture�asserts�that�if����g�cmmi12�A��is�a�mo�S�dular�ab�elian�v��X�ariet��ry�with��L�(�A;����1)�UR�!",�
cmsy10�6�=�0,�mthen��1H�������ō��[�L�(�A;����1)���[�[��z�%��
�΍�
�
������2cmmi8�A��������f�=������ō�l�#���hV1
wncyr10�X��
�n�(�A�)�������������u
cmex10�Q�����c�����p������[��z�qS �
�΍�#�A�(�&��N�cmbx12�Q�)�����|{Ycmr8�tor��������#�A�������K�cmsy8�_��*��(�Q�)�����tor������w��:��NF���d�Here�U��A�(�Q�)�����tor���
�is�the�group�of�rational�torsion�p�S�oin��rts�on��A�;��gthe�Shafarevic�h-T��Vate�group���X��M5�(�A�)�����dis���a�measure�of�the�failure�of�the�lo�S�cal-to-global�principle;�NTthe�T��Vamaga��rw�a���n�um�b�ers����c�����p��	�"�are�����dthe�
;orders�of�certain�comp�S�onen��rt�groups�asso�ciated�to��A�;�the�real�n��rum�b�er�
;
�����A��
��is�the�v��rolume�����dof�&�A�(�R�)�with�resp�S�ect�to�a�basis�of�dieren��rtials�ha�ving�ev�erywhere�nonzero�go�S�o�d�&reduction;�����dand���A����2�_��
X�is�the�dual�of��A�.������*�ǂ����6�;��䃍�������������	�Wil�8�liam��sA.�Stein�\'R���Uese�ar�ch��sSummary��(�����d��k�ff���������2ō��y�;����d�*��N�ffcmbx12�3.1��$��The�ffratio��+��g�ffcmmi12�L�)X�Qffcmr12�(�A;�fd�1)�=�
��(��
�b>

cmmi10�A���U���d�Extending���Manin's�w��rork�on�elliptic�curv�es,��A.�Agash���s�e�and�I���found�a�computable�form�ula������dfor��Qthe�rational�n��rum�b�S�er��Q�L�(�A;����1)�=�
�����A����.�e�Using�similar�tec��rhniques,�ҼI��"hop�e�to�nd�computable�����dform��rulas�#�for�rational�parts�of�sp�S�ecial�v��X�alues�of�t�wists,�K�and�of��L�-functions�attac�hed�to�forms�of�����dw��reigh�t���greater�than�t��rw�o.�3�I���ha�v�e���already�computed��L�(�A;����1)�=�
�����A��
���for�sev��reral�thousand�ab�S�elian�����dv��X�arieties,��and�hop�S�e�to�extend�these�computations.��"ʫ����d�3.2��$��The�ffT���famaga���w�a�n�um�b�s3ers��c��(��p������d�When�
��A��has�semistable�reduction�at��p�,��I�
�ha��rv�e�
�found�a�w��ra�y�
�to�explicitly�compute�the�n��rum-�����db�S�er���c�����p���]�,�up�to�a�p�o��rw�er��of�2.�8�I�hop�e�to�nd�a�w��ra�y��to�compute��c�����p����in�the�remaining�cases.��"ʫ����d�3.3��$��Bounding�ff�#���hV1ff
wncyr10�X������d�V.�l�Kolyv��X�agin�and�K.�Kato�obtained�upp�S�er�b�ounds�on�#��X��
�n�(�A�).�	��T��Vo�v��rerify�the�full�BSD�����dconjecture�n_for�certain�ab�S�elian�v��X�arieties,��Mit�is�necessary�is�to�mak��re�these�b�ounds�explicit.�����dKolyv��X�agin's�$b�S�ounds�in��rv�olv�e�$computations�with�Heegner�p�oin��rts,�2Xand�Kato's�in�v�olv�e�a�study�����dof��the�Galois�represen��rtations�asso�S�ciated�to��A�.��~I�Mplan�to�carry�out�suc�h�computations�in�man�y�����dsp�S�ecic��cases.�����XOne�>�approac��rh�to�sho�wing�that���X��6M�(�A�)�is�as�large�as�predicted�b�y�the�BSD�>�conjecture�is�����dsuggested��b��ry�Mazur's�notion�of�the�visible�part�of���X��ȋ�(�A�).��>Consider�an�ab�S�elian�v��X�ariet�y��A�����d�that�:�sits�naturally�in�the�Jacobian��J�����0����(�N�@�)�of�the�mo�S�dular�curv��re��X�����0���(�N�@�).�)�The��8���@cmti12�visible�}
p��ffart��of������d�X��
�ҹ(�A�)���is�the�collection�of�those�elemen��rts�of���X��y�(�A�)�that�go�to�0�under�the�natural�map�to������d�X��
�ҹ(�J�����0����(�N�@�)).�@�Cremona��Mand�Mazur�observ��red�that�if�an�elemen�t�of�order��p��in���X��仹(�A�)�is�visible,�����dthen�^Yit�is�explained�b��ry�a�jump�in�the�rank�of�Mordell-W��Veil,�zhin�the�sense�that�there�is�another�����dab�S�elian�eQsub��rv��X�ariet�y��B��X��UR�J�����0����(�N�@�)�suc�h�that��p�UR�j��#(�A��M�\��B���)�and��B�W�has�man��ry�rational�p�S�oin�ts.�nI�e/am�����dtrying���to�nd�the�precise�degree�to�whic��rh�this�observ��X�ation�can�b�S�e�turned�around:�u+if�there�����dis��another�ab�S�elian�v��X�ariet��ry��B��#�with�man�y�rational�p�S�oin�ts�and��p�Y��j��#(�A��T�\��B���),��then��under�what�����dh��ryp�S�otheses��is�there�an�elemen�t�of���X����(�A�)�of�order��p�?��(V����d�4��(XIcosahedral�z�Galois�represen��u�tations���#���d�E.���Artin�conjectured�that�the��L�-series�asso�S�ciated�to�an��ry�con�tin�uous�irreducible�represen�tation�����d��UR�:��G�����'2�@�cmbx8�Q��1��!���GL��������n���=�(�C�),�u�with�Xd�n�>��1,�is�Xden��rtire.�Recen�t�exciting�w�ork�of�T��Va�ylor�and�others�suggests�����dthat��a�complete�pro�S�of�of�Artin's�conjecture,�
1in�the�case�when��n��C�=�2��and����is�o�dd,�
1is�on�the�����dhorizon.�����XBy�IJcom��rbining�the�main�result�of�a�recen�t�pap�S�er�of�K.�Buzzard�and�T��Va�ylor�with�a�computer�����dcomputation,��~Buzzard�� and�I���recen��rtly�pro�v�ed�that�the�icosahedral�Artin�represen�tations�of�����dconductor���1376�UR=�2����2�5��l��Dh�43�are�mo�S�dular.�(0If�I���can�extend�a�congruence�result�of�J.�Sturm,�œthen�����dour�qGmetho�S�d�will�yield�sev��reral�more�examples.�jThese�ongoing�computations�are�la�ying�a�part�����dof��zthe�foundation�necessary�for�a�full�pro�S�of�of�the�Artin�conjecture�for�o�dd�t��rw�o-dimensional��z��,�����das�B�w��rell�as�stim�ulating�the�dev�elopmen�t�of�new�algorithms�for�computing�with�mo�S�dular�forms�����din��c��rharacteristic��`�.���������;��ܙ��	�8���@cmti12�7F
C��q
cmbxti10�+��g�ffcmmi12�*��N�ffcmbx12�)X�Qffcmr12�'2�@�cmbx8�&��N�cmbx12���u
cmex10��K�cmsy8�!",�
cmsy10��2cmmi8���g�cmmi12�|{Ycmr8���N�G�cmbx12���N��qcmbx12��hV1ff
wncyr10��hV1
wncyr10�X�Qcmr12�
�b>

cmmi10�!������