<meta http-equiv="Content-Type" content="text/html; charset=GB18030"><div style="font-family: Arial; width: 1146.65px; word-break: break-word !important;">Hi Adapters,</div><div style="font-family: Arial; width: 1146.65px; word-break: break-word !important;"><span style="font-family: arial; white-space: pre-wrap;"><br></span></div>
                
        
        
                <div style="width: 1146.65px; word-break: break-word !important;"><span style="font-family: NimbusRomNo9L; font-size: 10pt;">Math Word Problem (MWP) solving needs
to discover the quantitative relationships over
natural language narratives. Recent work
shows that existing models memorize proce-
dures from context and rely on shallow heuristics to solve MWPs.</span><span style="font-size: 10pt; white-space: pre-wrap;"><font face="arial"> In this talk, I will show some observations to answer the question "</font></span><font face="arial"><span style="font-size: 13.3333px; white-space: pre-wrap;">Are NLP Models really able to Solve Simple Math Word Problems?</span></font><span style="font-family: arial; font-size: 10pt; white-space: pre-wrap;">" which is also a NAACL paper in 2021. Inspired by these observations, I will also introduce a contrastive learning model for solving MWP.</span><span style="font-family: NimbusRomNo9L; font-size: 14pt;"></span></div>
                
        
        
                <div style="font-family: Arial; width: 1146.65px; word-break: break-word !important;"><span style="font-family: arial; white-space: pre-wrap;"><br></span></div><div style="font-family: Arial; width: 1146.65px; word-break: break-word !important;"><div dir="auto"><b style="word-break: break-word !important;">Time:</b> 2022/04/27 16:00-18:00</div><div dir="auto"><b style="word-break: break-word !important;">Venus:</b></div><div dir="auto">Tecent Meeting Link: <span style="  ; ; "> (</span><a href="https://meeting.tencent.com/dm/lvC8IXLVKApz" rel="noopener" target="_blank" style=" outline: none; cursor: pointer; color: rgb(0, 102, 153) ; ; ; ">https://meeting.tencent.com/dm/lvC8IXLVKApz</a><span style="  ; ; ">)</span></div><div dir="auto">Tecent Meeting Number£º<span style="  ; ; ">626-601-315</span></div><div dir="auto" style="word-break: break-word !important;"><div dir="auto" style="word-break: break-word !important;"><br></div><div dir="auto" style="word-break: break-word !important;">Best,</div></div></div><div style="font-family: Arial; width: 1146.65px; word-break: break-word !important;"><span style="font-family: arial; white-space: pre-wrap;">Shanshan</span></div>