From 9bfcf41b3c7551c6de9572fd491df2ecad63d8f2 Mon Sep 17 00:00:00 2001 From: Nadia DAOUDI Date: Thu, 21 Dec 2023 15:22:32 +0100 Subject: [PATCH] add seminar: 7 --- img/Xin-Cheng_Wen.jpg | Bin 0 -> 8528 bytes index.html | 64 ++++++++++++++++++++++++++++++------------ 2 files changed, 46 insertions(+), 18 deletions(-) create mode 100644 img/Xin-Cheng_Wen.jpg diff --git a/img/Xin-Cheng_Wen.jpg b/img/Xin-Cheng_Wen.jpg new file mode 100644 index 0000000000000000000000000000000000000000..309da6683739079a145969255c9288c40f2a5838 GIT binary patch literal 8528 zcmbVxd03Ozw{9pZPDmXP1(gaS3JR6cf|9n%REo%uL=rG{U~Wi}c}yJuMUeu96eyxh zDj^_45;B-n5Ktf@L&6Y3n1YNMK$1YF^LfvCe)pVv&mZUB{eA0s_LJ;y@BQwz_S)-R zTRp6v0Db+Fi@OU*Lqh}f3-Ex{lc00X31Pt?kf$fe3)v@$nX@ps<)&iyLSZNX-FV0NQJOdH9x^2g{ zty{Nm)zaFoy?y(Roj}prwo6BQ=dLgHFP(h({L5FsyK}47)-Nso--`Ml(C)3Ecc4N| z4I|K3yEQa-YpCCX3;_U}0jyu({zuXHN^{fZEx?qv?*JMQz6PeRsreN!+0C0a0j*Pk z|3RB}Z{BnC%nw`kdSBBrirQy&JNMaE;~yK|==$`Dj#*#7b!Xf5{dxxu>YE%lJ#o^^ z#@5dM>~{_qE;_roy1D=4>j#0tuK2@m+zbj12@MO6j){$nPe@FK_#ur=~AkBY=^)F=q2d>=!uCD+cY|{FIOXI6}Ks0x6+I;lPmOVdsYh8=l zYh-nM>%JdzpEbPMW^C;v(!GAGZ~OjZHj^geFVOxMvi}*dJO5wE{sruR;~E9+)YJf! zr@0#h1}PQgStmgMlKkIlz4j-_-9_hVKH)=biTBS_Hi-czg2!-b&@(Glj0V~5!+ttk z_C^ib{JYFR1P&bf%=8zK>}Z2XBj7Y{&oX9{8@Pa>Rx)wN?~el_->*% z9U|o3ulFXsemrSaOK;(Rzt`IhPbvI;`-$@ZTOwX%d+5ieSOc?ruLj-8jX;XLyhfg+ zdbFM4j3+#VyAlWw%CR3vz1^R1HGi>S;Lh(4F%-H&)eh{0*RZB1>zv7rR1@i2x4yXH zwu7rjVMuyyT3|OB&H#ZUgtF6}oxPkLe#BIC`;dMrzS!(Ey_7P3u;?JMtDDNtr3dl^ zL%oK3os1H4rjd^<3C!rrgM^2R^P{lj?u;jpV3lt0I+1OvFd`Y`$6#(h-7MC77|XbS zvYb1#+$6wAOtSYuhoRUS199xFQLZP%6`YiEUG}`P^&b}pv zLu$u5hKejlzOq*wVu|x5f#PD<>4*pCFq0)24@^m9RMY5NMq)X7Di_ONr$g3@=_h5` zBA3Nh$CPgWibSVbHAtWaCDA&e{o`y@UL=&O9DQcPex!PjnJDFn{E!)HkjUFI6S~ov z9_!(F-p7kk;T2OBsgV$N(VxeU9INTUQBkRO_!gL$qy}xu29%T;&JbM!kFH)6qq4@_ zD;M@um)4F{s$d*Em%eDnmg5PG30q}`%(&f-?kIYU#a|&61>xY%iJF2BGr9IUpT5t| z`fqGuw`v(`X;>Vnmmz!fjk+TU3<{{4*)*!<(xx4~%Vm6D8&4thGI|gb%5o z_sp0?c84hR1_KrtA{a93@&a_$_OZvA@Y}-=_AWIn$EiV%80z;X(OsB{GNkBI>IF#` z%Tf^dtZ5+T^^28XxwDR#?Fxbvdlxgo~Bt0Ms1EZG^9C{u#$ zJ6PeJT8nx9Ax_EV#bSiT^5|AIXnuD&ic6%!^!|cK7?i?HUeNlbBWGimuf=oA4lB+B zOSMyBD*BLf=u~m~V7Ih7WRxdcPSU!g22B9#s7=#}$@k|3ICPQ=4|004BNmxXEFUn1 zfQ}pf4aAvpXie6ahKfvhok3wsPl^(Xnf3!8g$~o|KhF#Js7A*NFyJ@qtJU z`l|tFc|Z2@_IAENxlqbf0)`n*$7%VqDh|HxP49Wtm0{#J-Kg~H4?j^(dHnUi{M*0# zIt@d?(CP!gJ?W*GJ~5P;tLBzz1>%Cj$Hclb^4Xx3QZ-0e$x(we>gQG%-(n`3kWI)| z#$zyn(OUm?_!FzlHYX=h&y$N$?h$}T+fJ=Jnm+s+qjS}=hl-WK8}Pn`r%NtZf3nG2 zO~ao1^w!V-AxRXaDX#+op$?Xr&8b?jR0C#d6^+*724*)_5BjZ_ z@W+^B98nyi*!Ljqz#K(KVKVAeWtTtIQ5fjp_pz^VihI~QKcti#5*#~W1 zPkQ8iI~y{$#y{W-2_WB?S6VOij!aA1MfJXcca~K zD?@KZ^Ba*^W!03|)=@#;_`x8LO~|G|Pb-UAQXHCF5xJg*rYaz-q)fmxNrh_A!BYyy zs)Xc8bbI?C?OH-H>d%eJudmG0qwRRb!e|TL>s2!~XovSDz(@xkbGex(`rNG3XR5)72Z)ZR4UV+0e4tN-2cCp~=t{r6}UrWP;@{LisvkWn1DRZ4c zkH3$)DSmI2RE1Jp9<;&SPCDP2MQE9(lz6!X#WeL{FVje_F`i5PbS}5+PTRXT5$c&oL3#X(oKAKI~?G*W}^Uc*;401AeMNYn_5AGkg50d_jD^{pK`J~U*a$F@>~tb1viqUysxvBh~Gbcw}Lr>MIzDN8lAJ&d!5cAqn0VnOX zS0nyp45>%VJCtHg5IxD5*QUBi_u|f#q(wX^;UlFb%6dtJM>`J`gW0LD?nipQYVyFP z;!nC^XU8~Y;cAc$bJaJ^2nkn%T7yhF8Zg?Wb&nsIzmIb0+#^sO4_=4K2?Gh;(|Q>( z!mmaMY(n*D+i@fJkRs#9E@%q1M9%Wgi-dg~$g8-V+Jnfs4Rtc>zd?43b3cYS;^&+j z>?I&cvC3j5P5IYnHK?icSYg_Y(YoPX?T{QL5GiMvheGg!Pbn~Pr_(VIaR#9J&A zM#I_nKN6*5_|;!nFJBH&)u7*R_&|<2nyxxID{IpXj$wALWgnTGBxj)P<_`}@vufu_ zXvOBZp#iYS9hq(i-lozD@Sk7)O*G~jDj0B|wVT{EPMG91Z(?uo+vb1e+-Qlt-IdCR zqKoPODDTs(4Kr2SM!OGO>ED3HW~VKkC^M>8IR^UwOcr~*x8FFDmTKZ4KsFm)EcO!K5RPccbV=5gh_g_DdCM+U}E>y?_rr7)*bgX7x8VI4Ii<@GF!=yJ3 zvxV%T;$&o4L)|c5;9r>qmO|(Po+6G0dwaxGmOZHk>7D|Mun09M_+9s_?rPjA%u}5zm%AZYSAERi@m40*ku3TZ7NLeG3SuNC%s51B+AISlr?;~3!x8QJG9g+#mq zU)Ra#Pk>GSUCiqLl_uUtPs&(4WWMrL4Z3}cCo+?)1rpYZF6sJ~SF-DoG{w-lP6t`h z2}bO%lrVN{pR?}bkrUp|`#}Fy?A-0l4F!F_IavT6hL{o|;@zoBJDlJk@4@Vm)l8`t zJ-5GsTqCCIqj*BJCqqC;SrlQXf@n;Me>DV6Efd1tmWiU!ihUZfIWjXwI^xX+p&EZp zg!arB`u0*I;(fAXGwDnu^0mBaj~$QF)vd5yLT*Vj31u`Jws8y(Cj2B0&T!_$MjA+o z)|h=h^&*lKA)~LQFuwlf(>9maui2zkr|I;D-32EvqsK4JCpsF2gx;B{mnfqM?_ccvG+s=h(+>x)y&#diZlOPJu?!YSKz-yT#gLonoRzpRa)XXE(g zO-WkyPb8&YbvX{uhvzYG%SMnowKt1qvQd-B>2R^-ePpxC-O1NZY7l~Dp?`R<^KGH< zkY!JUkpQEJg_FVi5+1%q?yE0a+g5CFMYR;R(y5ZlAuD zeVVg=d)xyl9u;p6^d@Lp9>poStbx1FPdQ)dzeA;*jbarXI``bZPsl3k6SgT~UPH9q z6}+%rjbZkM-+JAguJ}=mJE{iJ-SA!OTB@z?;!*JE4%IdHRQiDQx!_R~^T(9R1FyF0Q<{uyB$fqqeEZo?&k(+ z$e?hKj*S`=t0#0!K4B|^$&r{hAAp!brmFE#Q#yJH`B%g)Wcv@&V*C;*kEGbf;RC%!5Dy%D9rWznJLP~9 zh*KAG?x{gLYa@XH?>X#xDKt>Jq$`I7l%USCo@-WPs1V)$+2lgXP zB&eE7Mrz5R0z3ca`kk$8qj?PZ6ty@6DYW3j*Rk&XHy*bOA`c^b9wt>YafXU5O$yWJ zt}h~d-1>0kb*#;43qz=Lhm0vyllPy`{)L*!^BTA0j^R~nNg+#%a<>Gk1SxKexT-r? z*_2DH6uHrZ2ve9_e)3v|Oe*z{C}lTOZHJECF|-)xh2zR50WOfK+IV5$ns7682CC$s zKFK1&eA3-+`MKA+0V*mh|fJ{5M69*gx6smimXtHM2{kmA$bjb4-Ej{u#eF;yKrQ7og) zFsnYIN)ae}@>29c<=H^T2Y>SIk5_{dMdb-tm8PkjB{Q#_v#jR0^t<4WJ*wkba*>np zCS?}Z2Ng6-9^*Jc@e999+Q&taW5Eq&_Y4oy4vXcjmm|ivid{0NnEDez7N4na$vJZz zudZ<>Ggh9CG2m&Zq)zc}r+wq;`_K3(vEDsGDB{*wSp3Rp7Y{hCEQB)M)FQb9=FK}^ zv+y5dG-~RwudA$}>cYxj0N+C38EdOshPbN165Qizq^*dWed*I@ymJ~t)yWXWklyef zZP4^x-!gTeGT)a0*Y>OYG0irGXi1f`KM?8Ob|oomlqY4$V(T_FXg}=?x-J13(4X={ z#Z`+rbj)Hih4g2#!yZDrL-gI|zF_4(Zy&8ydth;aJr9uBln`aJ!aKG?X74{pKd3lA zsFNR2F&|xR%`Dokj7+vI}4W1uuzyvS|?C{1+teF!Rt7l`{*dS+dY6vOf33 zh-aO$xHs_^3of2mp8$3XylxPO;v6lIYV+DD^gE>o28KHzd!ZK#8-Hd~3ZF$uX9IC% z*lze^W(lGne~k-pxKC&i`}xBH*o7ipmOpH^{YJuJxaa;^PeyDCp;siRyK7(Rv9UJo zSx)5TRp`*Ok z9%^CD+?n++xhDVO^Y|QKmbGAELg@f&v7D2sX646s>B!q>BwaAcx@L*>^x;`Y-Cxpk4pLx++o4LDsSG?ouh z8BWzCd6tCjw>ky3y3*wGd;P{ce})8#=R9nuWJLT*>gqK|sVsqI|csoQ9OVeDV`#EK&_BOlAR5S{5Fxf90JdET+t{re2iF zE3O%H!NQXrGI)H;ZfRQh)<18)LqAIhPAws(E&;)vZv!2|)*qdZI>hUQzZ!FW8MMp|$xEk#zssLI>bC;NfPZ1g958K8CZdusHRwpQ0aFJ-_}zUMX;Ka)H?hb&RuP|K ztm4=GbIBb8D}BiH=YcJv>iXDCo9eQQMAqOn!K~?QyzxWlq;YiaU zPsSK@1R-IFJ?3F2S8B6=O|BQx8Y)(83smRkU zcq|kE$q3bY-TXn}tArYP_}FpwMGn<5jG=wazZrq&q^&fv&L@e{#zh0s=DQeea6@gW zYnUAm$OBv)Ox@k|;!5*qa+xFv6C0k7SZ6!I8|#Dx(TL*OHH!23ezIpZ;W7KF!;N()MGZ1rmB!U~U%~b%E7~kqL%MI4VuS^CSPyJZzc(-X zSQn#BVWS=DJ7^yeW^dNJnmP`Xv#{IO_t)ZAQJ!RNgkIjMKbXpPXUI{^Yvn>$3)14l zc}8@6k25GZH2U=GX>&8m`_&tA0(8TLFa%KrK4P?88e@d(s0K{6ePxX!g$BugaiUS{ z?I1K9ms$`{TqXw{Z^0#2pRtBGF4dMkU28SFa0c{jhFxQhqhj+Xxs;3`e>MP~nv9R(k zycB2N<}e*5%{Lu)K?5|#|2z25k{_fX$C>ijaA{-@n0hQZoOfqwwAT{`Qvt4y#_AYJ z02p3{|3O$hb7jc83P?vMIooi8C*-n@e!C( zBF?G~q2%M(@$N^;q72^Qf9n#wFchXPTrBvEh-)EGTsh=&=YKX*N5>j6&eEJ-Qe|-@ zrZhscmDE?q;;b}QPk^aJ(zAjaxdn%P`6veW>9$`d!;FJsp`O%RnAiUL_quoGF@yq$ zN`C;szcL`>@S(%J_A5y&Ha|9Klm`SBjZb%`{&@4T>&120Es||&y)mtym-NElt{!Y{ z!KzEX+@hQvXW092UM?rjwCbhtu(R!3QrN-h9DruaLd$j<9i}1(+1-Wkba^17IpwYfw`Ae`^ zSilh0^y}^&-bHdBJdWD$lAO)32cM3C2bdKnIWmpynF`Do?jBLpLx5i8pUg`aMt${{qa<7nIa5q!rrS$OP zWf)=2T|d7JyWwt9G%7}3YY(lyJJsENIGkInC{aKl6cbu|pTFxS z^JbyZt9HxG?++wpycimS_LVnX*!ixvOp06J8cs{!ATYAyscKN0yewmhchJK2U=T^X z9e$BXV!{v^WbSKSBWUK*0@c4HHU(Qj=`=7`2#?`Nstx!;^N4Qq!;HAmkF~GuEycF` zTXa+G@FpdZ9CF+Xz@_r>RD*-I**+gS*rOhc4@l+hdu*~zyLdn(mjiQr`hEtEYqCC7 zVjnV#rVt2dt1)+D;qb=_)E_B$s9QAyj%0*(R4&IibMVZ-rYhgS)!Q8Z~}QNJsz`e4ADNW@668?(O4FKW>EhkEZf68k|%}nRck&X?sTC04u^&-L(kUJ z;Qd1u&9cv&+3Y%t%9Wx#&vCzSoT!F3Y^{yJ>R#&umq%mECNPR1LC22vY_h)R>1! zJvhu*_=%@=8M`u=Q3++q(p$Kj7@Z!%fd9MY$}tq@`0{eul$^Qh?Yt$XhiQ6wWjc_9 zNhc#EP!h2Q-bF2dJWb`bt)j}aqQ*Wxv-t=Rb4rdSTFy&WChG$m*MVb1-&`|B$L#jYh#w;0IRpn-F~XBhCnAK sqdyWWUC~(&{?q%a-xHGAaeQ_yI^xiorqeXzP&FuY14t(Y8LCJA4HK62+yDRo literal 0 HcmV?d00001 diff --git a/index.html b/index.html index b476842..6c1e23f 100644 --- a/index.html +++ b/index.html @@ -177,23 +177,25 @@

Upcoming Seminars

- +
Zhong Li
-

Nanjing University

+

HIT

-

Robust Learning of Deep Predictive Models from Noisy and Imbalanced Software Engineering Datasets

-

With the rapid development of Deep Learning, deep predictive models have been widely applied to improve Software Engineering tasks, such as defect prediction and issue - classification, and have achieved remarkable success. They are mostly trained in a supervised manner, which heavily relies on high-quality datasets. Unfortunately, due to - the nature and source of software engineering data, the real-world datasets often suffer from the issues of sample mislabelling and class imbalance, thus undermining the - effectiveness of deep predictive models in practice. This problem has become a major obstacle for deep learning-based Software Engineering. In this paper, we propose - RobustTrainer, the first approach to learning deep predictive models on raw training datasets where the mislabelled samples and the imbalanced classes coexist. - RobustTrainer consists of a two-stage training scheme, where the first learns feature representations robust to sample mislabelling and the second builds a classifier robust - to class imbalance based on the learned representations in the first stage. We apply RobustTrainer to two popular Software Engineering tasks, i.e., Bug Report Classification - and Software Defect Prediction. Evaluation results show that RobustTrainer effectively tackles the mislabelling and class imbalance issues and produces significantly better - deep predictive models compared to the other six comparison approaches.

+

When Less is Enough: Positive and Unlabeled Learning Model for Vulnerability Detection

+

Automated code vulnerability detection has gained increasing attention in recent years. The deep learning (DL)-based methods, which implicitly learn vulnerable code patterns, have proven + effective in vulnerability detection. The performance of DL-based methods usually relies on the quantity and quality of labeled data. However, the current labeled data are generally automatically + collected, such as crawled from human-generated commits, making it hard to ensure the quality of the labels. Prior studies have demonstrated that the non-vulnerable code (i.e., negative labels) + tends to be unreliable in commonly-used datasets, while vulnerable code (i.e., positive labels) is more determined. Considering the large numbers of unlabeled data in practice, it is necessary and + worth exploring to leverage the positive data and large numbers of unlabeled data for more accurate vulnerability detection. In this paper, we focus on the Positive and Unlabeled (PU) learning problem + for vulnerability detection and propose a novel model named PILOT, i.e., Positive and unlabeled Learning mOdel for vulnerability deTection. PILOT only learns from positive and unlabeled data for + vulnerability detection. It mainly contains two modules: (1) A distance-aware label selection module, aiming at generating pseudo-labels for selected unlabeled data, which involves the inter-class + distance prototype and progressive fine-tuning; (2) A mixed-supervision representation learning module to further alleviate the influence of noise and enhance the discrimination of representations. + The experimental results show that PILOT outperforms + the popular weakly supervised methods by 2.78%-18.93% in the PU learning setting. Compared with the state-of-the-art methods, PILOT also improves the performance of 1.34%-12.46 % in F1 score metrics in + the supervised setting. In addition, PILOT can identify 23 mislabeled from the FFMPeg+Qemu dataset in the PU learning setting based on manual checking.

-

Presentation Date: Monday, December 18, 2023 at 10:30 AM CET

+

Presentation Date: Monday, January 29, 2024 at 10:30 AM CET

@@ -220,11 +222,40 @@

Past Seminars

  • + +
    + Robust Learning from Noisy and Imbalanced Software Engineering Datasets, Monday, December 4, 2023, by Zhong Li from NJU +
    +
    +
    +
    + +
    Zhong Li
    +

    Nanjing University

    +
    +
    +

    Robust Learning of Deep Predictive Models from Noisy and Imbalanced Software Engineering Datasets

    +

    With the rapid development of Deep Learning, deep predictive models have been widely applied to improve Software Engineering tasks, such as defect prediction and issue + classification, and have achieved remarkable success. They are mostly trained in a supervised manner, which heavily relies on high-quality datasets. Unfortunately, due to + the nature and source of software engineering data, the real-world datasets often suffer from the issues of sample mislabelling and class imbalance, thus undermining the + effectiveness of deep predictive models in practice. This problem has become a major obstacle for deep learning-based Software Engineering. In this paper, we propose + RobustTrainer, the first approach to learning deep predictive models on raw training datasets where the mislabelled samples and the imbalanced classes coexist. + RobustTrainer consists of a two-stage training scheme, where the first learns feature representations robust to sample mislabelling and the second builds a classifier robust + to class imbalance based on the learned representations in the first stage. We apply RobustTrainer to two popular Software Engineering tasks, i.e., Bug Report Classification + and Software Defect Prediction. Evaluation results show that RobustTrainer effectively tackles the mislabelling and class imbalance issues and produces significantly better + deep predictive models compared to the other six comparison approaches.

    + +

    Presentation Date: Monday, December 18, 2023 at 10:30 AM CET

    +
    +
    +
    +
    • + + +
  • Dataflow Analysis-Inspired DL for Efficient Vulnerability Detection, Monday, December 4, 2023, by Benjamin Steenhoek from ISU -
    -
    @@ -245,7 +276,6 @@

    Dataflow Analysis-Inspired Deep Learning for Efficient Vulnerability Detecti with 96.46 F1 score, 97.82 precision, and 95.14 recall.

    Presentation Date: Monday, December 4, 2023 at 3:00 PM CET

    -

    @@ -258,7 +288,6 @@

    Dataflow Analysis-Inspired Deep Learning for Efficient Vulnerability Detecti Towards Understanding Fairness and its Composition in Ensemble ML, Monday, November 20, 2023, by Usman Gohar from ISU -
    @@ -279,7 +308,6 @@

    Towards Understanding Fairness and its Composition in Ensemble Machine Learn ensemble design.

    Presentation Date: Monday, November 20, 2023 at 4:00 PM CET

    -