Overview of the NTCIR-14 FinNum Task Fine-Grained Numeral Understanding in Financial Social Media Data Chung-Chi Chen 1 , Hen-Hsen Huang 2,4 , Hiroya Takamura 3 , Hsin-Hsi Chen 1,4 1 Department of Computer Science and Information Engineering, National Taiwan University, Taiwan 2 Department of Computer Science, National Chengchi University, Taiwan 3 Artificial Intelligence Research Center, National Institute of Advanced Industrial Science and Technology, Japan 4 MOST Joint Research Center for AI Technology and All Vista Healthcare, Taiwan Fine-Grained Opinion on Social Trading Platforms Convert Crowd Opinions into Leading Indicator Numeral Taxonomy & FinNum Dataset Methods Experimental Results Error Analysis FinNum - 2 Numeral Attachment with Auxiliary Tasks. Numeral Attachment: $NE OK NE, last time oil was over $65 you were close to $8. Giddy-up… Semantic Understanding Features Representation Topic Format Position Keywords Named Entity Brown Cluster Part-of-speech Term Frequency Prefixes/ Suffixes Bag-of-Characters Numeral Information Recognizers-Text Type Models Skip-Gram GloVe ELMo BERT SVM MLP CNN RNN RNN + CNN Attention-based LSTM Task Setting Classification Sequential Labeling Further Directions Learning Numeracy Numeracy-600K: Learning Numeracy for Detecting Exaggerated Information in Market Comments Multilingual & Document Level Crowd View: Converting Investors' Opinions into Indicators. S&P 500 <.SPX> UP 1.53 POINTS, AT _____ DOW JONES<.DJI> UP 8.70 POINTS, AT _____ U.S. Q3 GDP rises _____ pct Participants