| { |
| "paper_id": "2022", |
| "header": { |
| "generated_with": "S2ORC 1.0.0", |
| "date_generated": "2023-01-19T12:16:52.379445Z" |
| }, |
| "title": "CMCL 2022 Shared Task on Multilingual and Crosslingual Prediction of Human Reading Behavior", |
| "authors": [ |
| { |
| "first": "Nora", |
| "middle": [], |
| "last": "Hollenstein", |
| "suffix": "", |
| "affiliation": {}, |
| "email": "nora.hollenstein@hum.ku.dk" |
| }, |
| { |
| "first": "Emmanuele", |
| "middle": [], |
| "last": "Chersoni", |
| "suffix": "", |
| "affiliation": {}, |
| "email": "emmanuelechersoni@gmail.com" |
| }, |
| { |
| "first": "Cassandra", |
| "middle": [], |
| "last": "Jacobs", |
| "suffix": "", |
| "affiliation": {}, |
| "email": "jacobs.cassandra.l@gmail.com" |
| }, |
| { |
| "first": "Yohei", |
| "middle": [], |
| "last": "Oseki", |
| "suffix": "", |
| "affiliation": {}, |
| "email": "oseki@g.ecc.u-tokyo.ac.jp" |
| }, |
| { |
| "first": "Laurent", |
| "middle": [], |
| "last": "Pr\u00e9vot", |
| "suffix": "", |
| "affiliation": {}, |
| "email": "laurent.prevot@univ-amu.fr" |
| }, |
| { |
| "first": "Enrico", |
| "middle": [], |
| "last": "Santus", |
| "suffix": "", |
| "affiliation": {}, |
| "email": "esantus@gmail.com" |
| } |
| ], |
| "year": "", |
| "venue": null, |
| "identifiers": {}, |
| "abstract": "We present the second shared task on eyetracking data prediction of the Cognitive Modeling and Computational Linguistics Workshop (CMCL). Differently from the previous edition, participating teams are asked to predict eyetracking features from multiple languages, including a surprise language for which there were no available training data. Moreover, the task also included the prediction of standard deviations of feature values in order to account for individual differences between readers. A total of six teams registered to the task. For the first subtask on multilingual prediction, the winning team proposed a regression model based on lexical features, while for the second subtask on cross-lingual prediction, the winning team used a hybrid model based on a multilingual transformer embeddings as well as statistical features.", |
| "pdf_parse": { |
| "paper_id": "2022", |
| "_pdf_hash": "", |
| "abstract": [ |
| { |
| "text": "We present the second shared task on eyetracking data prediction of the Cognitive Modeling and Computational Linguistics Workshop (CMCL). Differently from the previous edition, participating teams are asked to predict eyetracking features from multiple languages, including a surprise language for which there were no available training data. Moreover, the task also included the prediction of standard deviations of feature values in order to account for individual differences between readers. A total of six teams registered to the task. For the first subtask on multilingual prediction, the winning team proposed a regression model based on lexical features, while for the second subtask on cross-lingual prediction, the winning team used a hybrid model based on a multilingual transformer embeddings as well as statistical features.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Abstract", |
| "sec_num": null |
| } |
| ], |
| "body_text": [ |
| { |
| "text": "The benefits of eye movement data for machine learning have been assessed in various domains, including NLP (Barrett et al., 2016 (Barrett et al., , 2018 McGuire and Tomuro, 2021) and computer vision (Shanmuga Vadivel et al., 2015; Kruthiventi et al., 2017; Bautista and Naval, 2020; Tseng et al., 2020) . Eyetracking provides millisecond-accurate records on where humans look when they are reading and are useful in explanatory research of language processing. Eye movements depend on the stimulus and are therefore language-specific, but there are also universal tendencies that have been observed across languages (Liversedge et al., 2016) . Modelling human reading has been researched extensively in psycholinguistics (Reichle et al., 1998; Matthies and S\u00f8gaard, 2013; Hahn and Keller, 2016) . In NLP, eye-tracking prediction has been used to determine linguistic complexity (Singh et al., 2016; Sarti et al., 2021) or to analyze language models' ability to account for measures of human reading effort (Merkx and Frank) . Being able to accurately predict eye-tracking features across languages will advance this field and will facilitate comparisons between models and the analysis of their varying capabilities.", |
| "cite_spans": [ |
| { |
| "start": 108, |
| "end": 129, |
| "text": "(Barrett et al., 2016", |
| "ref_id": "BIBREF3" |
| }, |
| { |
| "start": 130, |
| "end": 153, |
| "text": "(Barrett et al., , 2018", |
| "ref_id": "BIBREF2" |
| }, |
| { |
| "start": 154, |
| "end": 179, |
| "text": "McGuire and Tomuro, 2021)", |
| "ref_id": "BIBREF37" |
| }, |
| { |
| "start": 200, |
| "end": 231, |
| "text": "(Shanmuga Vadivel et al., 2015;", |
| "ref_id": "BIBREF48" |
| }, |
| { |
| "start": 232, |
| "end": 257, |
| "text": "Kruthiventi et al., 2017;", |
| "ref_id": "BIBREF29" |
| }, |
| { |
| "start": 258, |
| "end": 283, |
| "text": "Bautista and Naval, 2020;", |
| "ref_id": "BIBREF4" |
| }, |
| { |
| "start": 284, |
| "end": 303, |
| "text": "Tseng et al., 2020)", |
| "ref_id": "BIBREF53" |
| }, |
| { |
| "start": 617, |
| "end": 642, |
| "text": "(Liversedge et al., 2016)", |
| "ref_id": "BIBREF34" |
| }, |
| { |
| "start": 722, |
| "end": 744, |
| "text": "(Reichle et al., 1998;", |
| "ref_id": "BIBREF44" |
| }, |
| { |
| "start": 745, |
| "end": 772, |
| "text": "Matthies and S\u00f8gaard, 2013;", |
| "ref_id": "BIBREF36" |
| }, |
| { |
| "start": 773, |
| "end": 795, |
| "text": "Hahn and Keller, 2016)", |
| "ref_id": "BIBREF16" |
| }, |
| { |
| "start": 879, |
| "end": 899, |
| "text": "(Singh et al., 2016;", |
| "ref_id": "BIBREF50" |
| }, |
| { |
| "start": 900, |
| "end": 919, |
| "text": "Sarti et al., 2021)", |
| "ref_id": "BIBREF47" |
| }, |
| { |
| "start": 1007, |
| "end": 1024, |
| "text": "(Merkx and Frank)", |
| "ref_id": null |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1" |
| }, |
| { |
| "text": "In this shared task, we address the challenge of predicting eye-tracking features recorded during sentence processing of multiple languages. We are interested in both cognitive modelling approaches as well as linguistically motivated approaches (i.e., language models). This shared task is hosted on CodaLab, where the instructions and pre-processed eye-tracking datasets are available. 1 Compared to the CMCL 2021 Shared Task on eye-tracking prediction (Hollenstein et al., 2021a) , we introduce two major changes:", |
| "cite_spans": [ |
| { |
| "start": 387, |
| "end": 388, |
| "text": "1", |
| "ref_id": null |
| }, |
| { |
| "start": 454, |
| "end": 481, |
| "text": "(Hollenstein et al., 2021a)", |
| "ref_id": "BIBREF21" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1" |
| }, |
| { |
| "text": "\u2022 Multilingual data: We provide an eye movement dataset with sentences from six different languages (Chinese, Dutch, English, German, Hindi, Russian) for Subtask 1 and a new Danish test set for Subtask 2.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1" |
| }, |
| { |
| "text": "\u2022 Eye-tracking features: To take into account the individual differences between readers, the task is not limited to predict the mean eye tracking features across readers, but also the standard deviation of the feature values.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1" |
| }, |
| { |
| "text": "2 Related Work", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1" |
| }, |
| { |
| "text": "It is widely acknowledged by researchers on naturalistic reading that fixation patterns are influenced by the words' contextual predictability (Ehrlich and Rayner, 1981) , although there is some substantial disagreement about the nature of this link (Brothers and Kuperberg, 2021). In Natural Language Processing, the most influential account of this phenomenon comes from surprisal theory (Hale, 2001; Levy, 2008) . This theory claims that the processing difficulty of a word is proportional to its surprisal, i.e., the negative logarithm of the probability of the word given the context, and it served as a reference framework for several studies on language models and eye-tracking data prediction (Demberg and Keller, 2008; Frank and Bod, 2011; Fossum and Levy, 2012) . Surprisal is not necessarily the only factor involved: for example, word length, word frequency, and other local statistics (e.g., bigram and trigram probabilities) also affect reading times (Rayner and Raney, 1996; Williams and Morris, 2004; Goodkind and Bicknell, 2021) . Embeddingbased semantic similarity was also found to be correlated with eye-tracking metrics (Mitchell et al., 2010; Salicchi et al., 2021; Yu et al., 2021) , although it is not clear whether its effect is independent of surprisal (Frank, 2017) . Later research work brought evidence that language models with a lower perplexity are better at fitting to human reading times (Goodkind and Bicknell, 2018; Aurnhammer and Frank, 2019; Wilcox et al., 2020; Merkx and Frank) . However, other studies suggested that perplexity may not tell the whole story. For example, Hao et al. (2020) pointed out that such a metric cannot be used for comparing models with different vocabularies and proposed, as a more reliable predictor, the correlation between surprisal values computed by a language model and the surprisal values obtained from humans by means of a Cloze test. Moreover, while most work on eye-tracking and language modeling focused on English, recent experiments on typologically distant languages like Japanese showed that lowerperplexity models may not be necessarily better at predicting eye-movement data (Kuribayashi et al., 2021) . Therefore, multilingual evaluation is an important step for building cognitively plausible models of human reading processes.", |
| "cite_spans": [ |
| { |
| "start": 143, |
| "end": 169, |
| "text": "(Ehrlich and Rayner, 1981)", |
| "ref_id": "BIBREF10" |
| }, |
| { |
| "start": 390, |
| "end": 402, |
| "text": "(Hale, 2001;", |
| "ref_id": "BIBREF17" |
| }, |
| { |
| "start": 403, |
| "end": 414, |
| "text": "Levy, 2008)", |
| "ref_id": "BIBREF33" |
| }, |
| { |
| "start": 701, |
| "end": 727, |
| "text": "(Demberg and Keller, 2008;", |
| "ref_id": "BIBREF8" |
| }, |
| { |
| "start": 728, |
| "end": 748, |
| "text": "Frank and Bod, 2011;", |
| "ref_id": "BIBREF13" |
| }, |
| { |
| "start": 749, |
| "end": 771, |
| "text": "Fossum and Levy, 2012)", |
| "ref_id": "BIBREF11" |
| }, |
| { |
| "start": 965, |
| "end": 989, |
| "text": "(Rayner and Raney, 1996;", |
| "ref_id": "BIBREF43" |
| }, |
| { |
| "start": 990, |
| "end": 1016, |
| "text": "Williams and Morris, 2004;", |
| "ref_id": "BIBREF55" |
| }, |
| { |
| "start": 1017, |
| "end": 1045, |
| "text": "Goodkind and Bicknell, 2021)", |
| "ref_id": "BIBREF15" |
| }, |
| { |
| "start": 1141, |
| "end": 1164, |
| "text": "(Mitchell et al., 2010;", |
| "ref_id": "BIBREF39" |
| }, |
| { |
| "start": 1165, |
| "end": 1187, |
| "text": "Salicchi et al., 2021;", |
| "ref_id": "BIBREF46" |
| }, |
| { |
| "start": 1188, |
| "end": 1204, |
| "text": "Yu et al., 2021)", |
| "ref_id": "BIBREF56" |
| }, |
| { |
| "start": 1279, |
| "end": 1292, |
| "text": "(Frank, 2017)", |
| "ref_id": "BIBREF12" |
| }, |
| { |
| "start": 1422, |
| "end": 1451, |
| "text": "(Goodkind and Bicknell, 2018;", |
| "ref_id": "BIBREF14" |
| }, |
| { |
| "start": 1452, |
| "end": 1479, |
| "text": "Aurnhammer and Frank, 2019;", |
| "ref_id": "BIBREF1" |
| }, |
| { |
| "start": 1480, |
| "end": 1500, |
| "text": "Wilcox et al., 2020;", |
| "ref_id": "BIBREF54" |
| }, |
| { |
| "start": 1501, |
| "end": 1517, |
| "text": "Merkx and Frank)", |
| "ref_id": "BIBREF38" |
| }, |
| { |
| "start": 1612, |
| "end": 1629, |
| "text": "Hao et al. (2020)", |
| "ref_id": "BIBREF19" |
| }, |
| { |
| "start": 2160, |
| "end": 2186, |
| "text": "(Kuribayashi et al., 2021)", |
| "ref_id": "BIBREF30" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Eye-Tracking and Language Models", |
| "sec_num": "2.1" |
| }, |
| { |
| "text": "Comparing monolingual and multilingual Transformer models, Hollenstein et al. (2021b) found that the latter are surprisingly accurate in predicting eye-tracking features across languages. In particular, multilingual BERT (Devlin et al., 2019) shows the best crosslinguistic transfer ability, even without being explicitly trained on the target language, while the XLM models (Lample and Conneau, 2019) achieve better in-language performance after fine-tuning.", |
| "cite_spans": [ |
| { |
| "start": 59, |
| "end": 85, |
| "text": "Hollenstein et al. (2021b)", |
| "ref_id": "BIBREF22" |
| }, |
| { |
| "start": 221, |
| "end": 242, |
| "text": "(Devlin et al., 2019)", |
| "ref_id": "BIBREF9" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Multilingual Eye-Tracking Corpora", |
| "sec_num": "2.2" |
| }, |
| { |
| "text": "Psycholinguistic research in the last two decades has led to the introduction of corpora with eyetracking recordings in several languages, including English (Cop et al., 2017; Luke and Christianson, 2017; Hollenstein et al., , 2020 , German (Kliegl et al., 2006; J\u00e4ger et al., 2021) , Hindi (Husain et al., 2015), Japanese (Asahara et al., 2016) , Dutch (Cop et al., 2017) , Russian (Laurinavichyute et al., 2019), Mandarin Chinese (Pan et al., 2021) , and Danish (Hollenstein et al., 2022) . However, it is not optimal to combine datasets recorded in different settings. The most recent release is the Multilingual Eye-Movement Corpus (MECO; Siegelman et al. 2022), a new resource including parallel data from 580 readers of 13 different languages following the same experiment protocol. The notable differences between these corpora and other psycholinguistic studies is the naturally occurring stimuli, the presentation of full sentences or longer text spans, and that the participants were able to read in their own speed. ", |
| "cite_spans": [ |
| { |
| "start": 157, |
| "end": 175, |
| "text": "(Cop et al., 2017;", |
| "ref_id": "BIBREF7" |
| }, |
| { |
| "start": 176, |
| "end": 204, |
| "text": "Luke and Christianson, 2017;", |
| "ref_id": "BIBREF35" |
| }, |
| { |
| "start": 205, |
| "end": 231, |
| "text": "Hollenstein et al., , 2020", |
| "ref_id": "BIBREF24" |
| }, |
| { |
| "start": 241, |
| "end": 262, |
| "text": "(Kliegl et al., 2006;", |
| "ref_id": "BIBREF28" |
| }, |
| { |
| "start": 263, |
| "end": 282, |
| "text": "J\u00e4ger et al., 2021)", |
| "ref_id": "BIBREF27" |
| }, |
| { |
| "start": 323, |
| "end": 345, |
| "text": "(Asahara et al., 2016)", |
| "ref_id": "BIBREF0" |
| }, |
| { |
| "start": 354, |
| "end": 372, |
| "text": "(Cop et al., 2017)", |
| "ref_id": "BIBREF7" |
| }, |
| { |
| "start": 432, |
| "end": 450, |
| "text": "(Pan et al., 2021)", |
| "ref_id": "BIBREF40" |
| }, |
| { |
| "start": 464, |
| "end": 490, |
| "text": "(Hollenstein et al., 2022)", |
| "ref_id": "BIBREF20" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Multilingual Eye-Tracking Corpora", |
| "sec_num": "2.2" |
| }, |
| { |
| "text": "The shared task is formulated as a regression task to predict 2 eye-tracking features and the corresponding standard deviation across readers for each word:", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Task Description", |
| "sec_num": "3" |
| }, |
| { |
| "text": "1. FFDAVG: first fixation duration (FFD), the duration of the first fixation on the prevailing word;", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Task Description", |
| "sec_num": "3" |
| }, |
| { |
| "text": "2. FFDSTD: standard deviation of FFD across readers;", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Task Description", |
| "sec_num": "3" |
| }, |
| { |
| "text": "3. TRTAVG: total reading time (TRT), the sum of all fixation durations on the current word, including regressions; 4. TRTSTD: standard deviation of TRT across readers.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Task Description", |
| "sec_num": "3" |
| }, |
| { |
| "text": "The goal of the first subtask is multilingual eye tracking prediction, i.e., to predict the eye-tracking features for sentences of the 6 provided languages in the training data on held-out sentences of the same languages in the test data. The dataset contains sentences from a range of openly available eye-tracking corpora.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Subtask 1", |
| "sec_num": "3.1" |
| }, |
| { |
| "text": "The second subtask test the models' performances in a cross-lingual prediction scenario. The training and development data are identical to Subtask 1, but the test data contains eye-tracking data from a new language. The participants were only informed about which language would be included in this subtask at the beginning of the evaluation phase. sentences read by 5 readers, extracted from the CopCo corpus (Hollenstein et al., 2022) .", |
| "cite_spans": [ |
| { |
| "start": 411, |
| "end": 437, |
| "text": "(Hollenstein et al., 2022)", |
| "ref_id": "BIBREF20" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Subtask 2", |
| "sec_num": "3.2" |
| }, |
| { |
| "text": "Tokenization The tokens in the sentences are split in the same manner as they were presented to the participants during the reading experiments. Hence, this does not necessarily follow a linguistically correct tokenization. For example, the sequences \"(except,\" and \"don't\" were presented as such to the reader and not split into \"(\", \"except\", \",\" and \"do\", \"n't\" as a tokenizer would do. It is the participants' decision how to deal with these tokens.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Preprocessing", |
| "sec_num": "4.3" |
| }, |
| { |
| "text": "Feature Extraction The data contains scaled features in the range between 0 and 100 to facilitate evaluation via the mean absolute average (MAE). The eye-tracking feature values (FFDAVG and TR-TAVG) are averaged over all available readers of a corpus. This preprocessing step is done separately for each corpus before combining them. Table 2 shows the scaled features values across the full dataset of Subtask 1. In Figure 2 , we present the distributions of the feature values for the training set and the test set of Subtasks 1 and 2. Finally, Figure 3 in the Appendix shows the individual plots for each language.", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 334, |
| "end": 341, |
| "text": "Table 2", |
| "ref_id": "TABREF2" |
| }, |
| { |
| "start": 416, |
| "end": 424, |
| "text": "Figure 2", |
| "ref_id": null |
| }, |
| { |
| "start": 546, |
| "end": 554, |
| "text": "Figure 3", |
| "ref_id": null |
| } |
| ], |
| "eq_spans": [], |
| "section": "Preprocessing", |
| "sec_num": "4.3" |
| }, |
| { |
| "text": "In this section, we describe the evaluation procedure used to assess the submitted predictions of the participating teams. Any additional data source was allowed to train the models, as long as it is freely available to the research community. For example, additional eyetracking corpora, additional features such as brain activity signals, pre-trained language models, etc.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Evaluation", |
| "sec_num": "5" |
| }, |
| { |
| "text": "The submitted predictions are evaluated against the real eye-tracking feature values using the mean absolute error (MAE) metric, a measure of errors between paired observations including comparisons of predicted (y) versus observed (x) values for each word in the test set:", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Scoring Metric", |
| "sec_num": "5.1" |
| }, |
| { |
| "text": "M AE = n i=1 |y i \u2212 x i | n (1)", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Scoring Metric", |
| "sec_num": "5.1" |
| }, |
| { |
| "text": "The winning system is defined as the one with the lowest average MAE across all 4 features. We reported additional metrics for analysis, namely R 2 for all features individually and aggregated, but only MAE was used for the ranking.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Scoring Metric", |
| "sec_num": "5.1" |
| }, |
| { |
| "text": "We use the mean central tendency as a baseline for this regression problem, i.e., we calculate the mean value for each feature from the training data and use it as a prediction for all words in the test data. Table 3 shows the MAE scores achieved by this mean baseline for each eye-tracking feature. For Subtask 1, we add an additional stronger mean baseline calculated over the training set of each language individually. This baseline assumes that the language of each sentence is known to the system. This second baseline was not reported in the rankings, but serves for further analysis.", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 209, |
| "end": 216, |
| "text": "Table 3", |
| "ref_id": "TABREF5" |
| } |
| ], |
| "eq_spans": [], |
| "section": "Mean Baseline", |
| "sec_num": "5.2" |
| }, |
| { |
| "text": "Six teams and a total of 37 participants registered on the competition website. All six teams submitted their predictions during the evaluation phase for Subtask 1. Four of the teams also submitted predictions for Subtask 2. Each team was allowed three submissions for each subtask during the evaluation phases. Finally, 5 teams published system description papers outlining their approaches (see Table 3 for all references).", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 397, |
| "end": 404, |
| "text": "Table 3", |
| "ref_id": "TABREF5" |
| } |
| ], |
| "eq_spans": [], |
| "section": "Participating Teams & Systems", |
| "sec_num": "6" |
| }, |
| { |
| "text": "The participating teams submitted predictions generated from various approaches, from regression algorithms such as random forests (NU HLT) and linear regression models (HkAmsters) with a wide range of lexical, cognitively and phoneticallymotivated features (NU HLT), to neural approaches that fine-tune large pre-trained transformer models with additional regression heads, and integrate adapters into pre-trained transformer language models (DMG) (Pfeiffer et al., 2020; Han et al., 2021) . Some teams chose to build language-specific models (e.g., HkAmsters, DMG), while others merged the words from all languages into a common vocabulary space in which all words are converted to their IPA forms (NU HLT). Moreover, representations from both monolingual models such as GPT-2 (Radford et al., 2019) as well as multilingual transformer models such as mBERT (Devlin et al., 2019) and XLM (Lample and Conneau, 2019) were also included (Poirot). For the second subtask, dealing with a new unseen language was handled again through a common phonetic vocabulary space (NU HLT), through translation (i.e., translating the Danish text to German and then using a German model for prediction) (DMG), or zero-shot learning (Poirot).", |
| "cite_spans": [ |
| { |
| "start": 449, |
| "end": 472, |
| "text": "(Pfeiffer et al., 2020;", |
| "ref_id": "BIBREF41" |
| }, |
| { |
| "start": 473, |
| "end": 490, |
| "text": "Han et al., 2021)", |
| "ref_id": "BIBREF18" |
| }, |
| { |
| "start": 779, |
| "end": 801, |
| "text": "(Radford et al., 2019)", |
| "ref_id": "BIBREF42" |
| }, |
| { |
| "start": 859, |
| "end": 880, |
| "text": "(Devlin et al., 2019)", |
| "ref_id": "BIBREF9" |
| }, |
| { |
| "start": 889, |
| "end": 915, |
| "text": "(Lample and Conneau, 2019)", |
| "ref_id": "BIBREF31" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Methods", |
| "sec_num": "6.1" |
| }, |
| { |
| "text": "In this section, we describe the prediction performance achieved by the participating teams. The official results of this shared task are presented in Tables 3 and 4 for Subtask 1 and 2, respectively. The best results for the first subtask on multilingual prediction were achieved by Team HkAmsters with language-specific regression models based on word-level features such as word length, word frequency, and surprisal scores estimated with GPT-2 (Radford et al., 2019) . For the second subtask on cross-lingual prediction, the winning team (Poirot) used a zero-shot hybrid model based on a multilingual transformer embeddings as well as statistical features.", |
| "cite_spans": [ |
| { |
| "start": 448, |
| "end": 470, |
| "text": "(Radford et al., 2019)", |
| "ref_id": "BIBREF42" |
| } |
| ], |
| "ref_spans": [ |
| { |
| "start": 151, |
| "end": 165, |
| "text": "Tables 3 and 4", |
| "ref_id": "TABREF5" |
| } |
| ], |
| "eq_spans": [], |
| "section": "Results", |
| "sec_num": "7" |
| }, |
| { |
| "text": "We presented the results of the second shared task on predicting token-level eye-tracking features recorded during natural reading of sentences or longer text spans. In this second edition, we focused on multilingual and crosslingual prediction. We hope the CMCL Shared Task makes a lasting contribution to the field of linguistic cognitive modelling by providing researchers with a standard evaluation framework and a high quality dataset. Despite the limited size of the training and test sets as well as the diversity of text domains within the eye-tracking corpora, many previously reached conclusions can now be tested more thoroughly and future models can be compared on a shared multilingual benchmark.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Outlook & Conclusion", |
| "sec_num": "8" |
| }, |
| { |
| "text": "A Appendix Figure 3 shows the distributions of the feature values for the data of all six languages. Figure 3 : Boxplots showing the feature value distributions of the eye-tracking data of all languages of Subtask 1. Below each box is the median value of each feature.", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 11, |
| "end": 19, |
| "text": "Figure 3", |
| "ref_id": null |
| }, |
| { |
| "start": 101, |
| "end": 109, |
| "text": "Figure 3", |
| "ref_id": null |
| } |
| ], |
| "eq_spans": [], |
| "section": "Outlook & Conclusion", |
| "sec_num": "8" |
| }, |
| { |
| "text": "https://competitions.codalab.org/ competitions/36415", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "", |
| "sec_num": null |
| } |
| ], |
| "back_matter": [ |
| { |
| "text": "Emmanuele Chersoni acknowledges the Startup Fund for RAPs under the Strategic Hiring Scheme from the Hong Kong Polytechnic University (PolyU UGC, BD8S).", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Acknowledgements", |
| "sec_num": null |
| } |
| ], |
| "bib_entries": { |
| "BIBREF0": { |
| "ref_id": "b0", |
| "title": "Reading-Time Annotations for \"Balanced Corpus of Contemporary Written Japanese", |
| "authors": [ |
| { |
| "first": "Masayuki", |
| "middle": [], |
| "last": "Asahara", |
| "suffix": "" |
| }, |
| { |
| "first": "Hajime", |
| "middle": [], |
| "last": "Ono", |
| "suffix": "" |
| }, |
| { |
| "first": "Edson T", |
| "middle": [], |
| "last": "Miyamoto", |
| "suffix": "" |
| } |
| ], |
| "year": 2016, |
| "venue": "Proceedings of COLING: Technical Papers", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Masayuki Asahara, Hajime Ono, and Edson T Miyamoto. 2016. Reading-Time Annotations for \"Balanced Corpus of Contemporary Written Japanese\". In Proceedings of COLING: Technical Papers.", |
| "links": null |
| }, |
| "BIBREF1": { |
| "ref_id": "b1", |
| "title": "Evaluating Information-theoretic Measures of Word Prediction in Naturalistic Sentence Reading", |
| "authors": [ |
| { |
| "first": "Christoph", |
| "middle": [], |
| "last": "Aurnhammer", |
| "suffix": "" |
| }, |
| { |
| "first": "L", |
| "middle": [], |
| "last": "Stefan", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Frank", |
| "suffix": "" |
| } |
| ], |
| "year": 2019, |
| "venue": "Neuropsychologia", |
| "volume": "134", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Christoph Aurnhammer and Stefan L Frank. 2019. Eval- uating Information-theoretic Measures of Word Pre- diction in Naturalistic Sentence Reading. Neuropsy- chologia, 134:107198.", |
| "links": null |
| }, |
| "BIBREF2": { |
| "ref_id": "b2", |
| "title": "Sequence Classification with Human Attention", |
| "authors": [ |
| { |
| "first": "Maria", |
| "middle": [], |
| "last": "Barrett", |
| "suffix": "" |
| }, |
| { |
| "first": "Joachim", |
| "middle": [], |
| "last": "Bingel", |
| "suffix": "" |
| }, |
| { |
| "first": "Nora", |
| "middle": [], |
| "last": "Hollenstein", |
| "suffix": "" |
| }, |
| { |
| "first": "Marek", |
| "middle": [], |
| "last": "Rei", |
| "suffix": "" |
| }, |
| { |
| "first": "Anders", |
| "middle": [], |
| "last": "S\u00f8gaard", |
| "suffix": "" |
| } |
| ], |
| "year": 2018, |
| "venue": "Proceedings of the 22nd Conference on Computational Natural Language Learning", |
| "volume": "", |
| "issue": "", |
| "pages": "302--312", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Maria Barrett, Joachim Bingel, Nora Hollenstein, Marek Rei, and Anders S\u00f8gaard. 2018. Sequence Classifi- cation with Human Attention. In Proceedings of the 22nd Conference on Computational Natural Lan- guage Learning, pages 302-312.", |
| "links": null |
| }, |
| "BIBREF3": { |
| "ref_id": "b3", |
| "title": "Weakly Supervised Part-ofspeech Tagging Using Eye-tracking Data", |
| "authors": [ |
| { |
| "first": "Maria", |
| "middle": [], |
| "last": "Barrett", |
| "suffix": "" |
| }, |
| { |
| "first": "Joachim", |
| "middle": [], |
| "last": "Bingel", |
| "suffix": "" |
| }, |
| { |
| "first": "Frank", |
| "middle": [], |
| "last": "Keller", |
| "suffix": "" |
| }, |
| { |
| "first": "Anders", |
| "middle": [], |
| "last": "S\u00f8gaard", |
| "suffix": "" |
| } |
| ], |
| "year": 2016, |
| "venue": "Proceedings of ACL", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Maria Barrett, Joachim Bingel, Frank Keller, and An- ders S\u00f8gaard. 2016. Weakly Supervised Part-of- speech Tagging Using Eye-tracking Data. In Pro- ceedings of ACL.", |
| "links": null |
| }, |
| "BIBREF4": { |
| "ref_id": "b4", |
| "title": "Towards Learning to Read Like Humans", |
| "authors": [ |
| { |
| "first": "Louise", |
| "middle": [ |
| "Gillian" |
| ], |
| "last": "Bautista", |
| "suffix": "" |
| }, |
| { |
| "first": "Prospero", |
| "middle": [], |
| "last": "Naval", |
| "suffix": "" |
| } |
| ], |
| "year": 2020, |
| "venue": "International Conference on Computational Collective Intelligence", |
| "volume": "", |
| "issue": "", |
| "pages": "779--791", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Louise Gillian Bautista and Prospero Naval. 2020. To- wards Learning to Read Like Humans. In Interna- tional Conference on Computational Collective Intel- ligence, pages 779-791. Springer.", |
| "links": null |
| }, |
| "BIBREF5": { |
| "ref_id": "b5", |
| "title": "Team \u00daFAL at CMCL 2022 Shared Task: Figuring Out the Correct Recipe for Predicting Eye-Tracking Features Using Pretrained Language Models", |
| "authors": [ |
| { |
| "first": "Sunit", |
| "middle": [], |
| "last": "Bhattacharya", |
| "suffix": "" |
| }, |
| { |
| "first": "Rishu", |
| "middle": [], |
| "last": "Kumar", |
| "suffix": "" |
| }, |
| { |
| "first": "Ondrej", |
| "middle": [], |
| "last": "Bojar", |
| "suffix": "" |
| } |
| ], |
| "year": 2022, |
| "venue": "Proceedings of the ACL Workshop on Cognitive Modeling and Computational Linguistics", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Sunit Bhattacharya, Rishu Kumar, and Ondrej Bojar. 2022. Team \u00daFAL at CMCL 2022 Shared Task: Figuring Out the Correct Recipe for Predicting Eye- Tracking Features Using Pretrained Language Mod- els. In Proceedings of the ACL Workshop on Cogni- tive Modeling and Computational Linguistics.", |
| "links": null |
| }, |
| "BIBREF6": { |
| "ref_id": "b6", |
| "title": "Word Predictability Effects Are Linear, Not Logarithmic: Implications for Probabilistic Models of Sentence Comprehension", |
| "authors": [ |
| { |
| "first": "Trevor", |
| "middle": [], |
| "last": "Brothers", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Gina R Kuperberg", |
| "suffix": "" |
| } |
| ], |
| "year": 2021, |
| "venue": "Journal of Memory and Language", |
| "volume": "116", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Trevor Brothers and Gina R Kuperberg. 2021. Word Predictability Effects Are Linear, Not Logarithmic: Implications for Probabilistic Models of Sentence Comprehension. Journal of Memory and Language, 116:104174.", |
| "links": null |
| }, |
| "BIBREF7": { |
| "ref_id": "b7", |
| "title": "Presenting GECO: An Eyetracking Corpus of Monolingual and Bilingual Sentence Reading", |
| "authors": [ |
| { |
| "first": "Uschi", |
| "middle": [], |
| "last": "Cop", |
| "suffix": "" |
| }, |
| { |
| "first": "Nicolas", |
| "middle": [], |
| "last": "Dirix", |
| "suffix": "" |
| }, |
| { |
| "first": "Denis", |
| "middle": [], |
| "last": "Drieghe", |
| "suffix": "" |
| }, |
| { |
| "first": "Wouter", |
| "middle": [], |
| "last": "Duyck", |
| "suffix": "" |
| } |
| ], |
| "year": 2017, |
| "venue": "Behavior Research Methods", |
| "volume": "49", |
| "issue": "2", |
| "pages": "602--615", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Uschi Cop, Nicolas Dirix, Denis Drieghe, and Wouter Duyck. 2017. Presenting GECO: An Eyetracking Corpus of Monolingual and Bilingual Sentence Read- ing. Behavior Research Methods, 49(2):602-615.", |
| "links": null |
| }, |
| "BIBREF8": { |
| "ref_id": "b8", |
| "title": "Data from Eyetracking Corpora as Evidence for Theories of Syntactic Processing Complexity", |
| "authors": [ |
| { |
| "first": "Vera", |
| "middle": [], |
| "last": "Demberg", |
| "suffix": "" |
| }, |
| { |
| "first": "Frank", |
| "middle": [], |
| "last": "Keller", |
| "suffix": "" |
| } |
| ], |
| "year": 2008, |
| "venue": "Cognition", |
| "volume": "109", |
| "issue": "2", |
| "pages": "193--210", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Vera Demberg and Frank Keller. 2008. Data from Eye- tracking Corpora as Evidence for Theories of Syntac- tic Processing Complexity. Cognition, 109(2):193- 210.", |
| "links": null |
| }, |
| "BIBREF9": { |
| "ref_id": "b9", |
| "title": "BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding", |
| "authors": [ |
| { |
| "first": "Jacob", |
| "middle": [], |
| "last": "Devlin", |
| "suffix": "" |
| }, |
| { |
| "first": "Ming-Wei", |
| "middle": [], |
| "last": "Chang", |
| "suffix": "" |
| }, |
| { |
| "first": "Kenton", |
| "middle": [], |
| "last": "Lee", |
| "suffix": "" |
| }, |
| { |
| "first": "Kristina", |
| "middle": [], |
| "last": "Toutanova", |
| "suffix": "" |
| } |
| ], |
| "year": 2019, |
| "venue": "Proceedings of NAACL", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. BERT: Pre-training of Deep Bidirectional Transformers for Language Un- derstanding. In Proceedings of NAACL.", |
| "links": null |
| }, |
| "BIBREF10": { |
| "ref_id": "b10", |
| "title": "Contextual Effects on Word Perception and Eye Movements During Reading", |
| "authors": [ |
| { |
| "first": "Susan", |
| "middle": [ |
| "E" |
| ], |
| "last": "Ehrlich", |
| "suffix": "" |
| }, |
| { |
| "first": "Keith", |
| "middle": [], |
| "last": "Rayner", |
| "suffix": "" |
| } |
| ], |
| "year": 1981, |
| "venue": "Journal of Verbal Learning and Verbal Behavior", |
| "volume": "20", |
| "issue": "", |
| "pages": "641--65", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Susan E. Ehrlich and Keith Rayner. 1981. Contextual Effects on Word Perception and Eye Movements Dur- ing Reading. Journal of Verbal Learning and Verbal Behavior, 20:641-65.", |
| "links": null |
| }, |
| "BIBREF11": { |
| "ref_id": "b11", |
| "title": "Sequential vs. Hierarchical Syntactic Models of Human Incremental Sentence Processing", |
| "authors": [ |
| { |
| "first": "Victoria", |
| "middle": [], |
| "last": "Fossum", |
| "suffix": "" |
| }, |
| { |
| "first": "Roger", |
| "middle": [], |
| "last": "Levy", |
| "suffix": "" |
| } |
| ], |
| "year": 2012, |
| "venue": "Proceedings of the NAACL Workshop on Cognitive Modeling and Computational Linguistics", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Victoria Fossum and Roger Levy. 2012. Sequential vs. Hierarchical Syntactic Models of Human Incremental Sentence Processing. In Proceedings of the NAACL Workshop on Cognitive Modeling and Computational Linguistics.", |
| "links": null |
| }, |
| "BIBREF12": { |
| "ref_id": "b12", |
| "title": "Word Embedding Distance Does Not Predict Word Reading Time", |
| "authors": [ |
| { |
| "first": "L", |
| "middle": [], |
| "last": "Stefan", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Frank", |
| "suffix": "" |
| } |
| ], |
| "year": 2017, |
| "venue": "Proceedings of CogSci", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Stefan L Frank. 2017. Word Embedding Distance Does Not Predict Word Reading Time. In Proceedings of CogSci.", |
| "links": null |
| }, |
| "BIBREF13": { |
| "ref_id": "b13", |
| "title": "Insensitivity of the Human Sentence-processing System to Hierarchical Structure", |
| "authors": [ |
| { |
| "first": "L", |
| "middle": [], |
| "last": "Stefan", |
| "suffix": "" |
| }, |
| { |
| "first": "Rens", |
| "middle": [], |
| "last": "Frank", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Bod", |
| "suffix": "" |
| } |
| ], |
| "year": 2011, |
| "venue": "Psychological Science", |
| "volume": "22", |
| "issue": "6", |
| "pages": "829--834", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Stefan L Frank and Rens Bod. 2011. Insensitivity of the Human Sentence-processing System to Hierarchical Structure. Psychological Science, 22(6):829-834.", |
| "links": null |
| }, |
| "BIBREF14": { |
| "ref_id": "b14", |
| "title": "Predictive Power of Word Surprisal for Reading Times Is a Linear Function of Language Model Quality", |
| "authors": [ |
| { |
| "first": "Adam", |
| "middle": [], |
| "last": "Goodkind", |
| "suffix": "" |
| }, |
| { |
| "first": "Klinton", |
| "middle": [], |
| "last": "Bicknell", |
| "suffix": "" |
| } |
| ], |
| "year": 2018, |
| "venue": "Proceedings of the LSA Workshop on Cognitive Modeling and Computational Linguistics", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Adam Goodkind and Klinton Bicknell. 2018. Predic- tive Power of Word Surprisal for Reading Times Is a Linear Function of Language Model Quality. In Proceedings of the LSA Workshop on Cognitive Mod- eling and Computational Linguistics.", |
| "links": null |
| }, |
| "BIBREF15": { |
| "ref_id": "b15", |
| "title": "Local Word Statistics Affect Reading Times Independently of Surprisal", |
| "authors": [ |
| { |
| "first": "Adam", |
| "middle": [], |
| "last": "Goodkind", |
| "suffix": "" |
| }, |
| { |
| "first": "Klinton", |
| "middle": [], |
| "last": "Bicknell", |
| "suffix": "" |
| } |
| ], |
| "year": 2021, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": { |
| "arXiv": [ |
| "arXiv:2103.04469" |
| ] |
| }, |
| "num": null, |
| "urls": [], |
| "raw_text": "Adam Goodkind and Klinton Bicknell. 2021. Local Word Statistics Affect Reading Times Independently of Surprisal. arXiv preprint arXiv:2103.04469.", |
| "links": null |
| }, |
| "BIBREF16": { |
| "ref_id": "b16", |
| "title": "Modeling Human Reading with Neural Attention", |
| "authors": [ |
| { |
| "first": "Michael", |
| "middle": [], |
| "last": "Hahn", |
| "suffix": "" |
| }, |
| { |
| "first": "Frank", |
| "middle": [], |
| "last": "Keller", |
| "suffix": "" |
| } |
| ], |
| "year": 2016, |
| "venue": "Proceedings of EMNLP", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Michael Hahn and Frank Keller. 2016. Modeling Hu- man Reading with Neural Attention. In Proceedings of EMNLP.", |
| "links": null |
| }, |
| "BIBREF17": { |
| "ref_id": "b17", |
| "title": "A Probabilistic Earley Parser as a Psycholinguistic Model", |
| "authors": [ |
| { |
| "first": "John", |
| "middle": [], |
| "last": "Hale", |
| "suffix": "" |
| } |
| ], |
| "year": 2001, |
| "venue": "Proceedings of NAACL", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "John Hale. 2001. A Probabilistic Earley Parser as a Psycholinguistic Model. In Proceedings of NAACL.", |
| "links": null |
| }, |
| "BIBREF18": { |
| "ref_id": "b18", |
| "title": "Robust Transfer Learning with Pretrained Language Models through Adapters", |
| "authors": [ |
| { |
| "first": "Wenjuan", |
| "middle": [], |
| "last": "Han", |
| "suffix": "" |
| }, |
| { |
| "first": "Bo", |
| "middle": [], |
| "last": "Pang", |
| "suffix": "" |
| }, |
| { |
| "first": "Yingnian", |
| "middle": [], |
| "last": "Wu", |
| "suffix": "" |
| } |
| ], |
| "year": 2021, |
| "venue": "Proceedings of ACL", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Wenjuan Han, Bo Pang, and Yingnian Wu. 2021. Ro- bust Transfer Learning with Pretrained Language Models through Adapters. In Proceedings of ACL.", |
| "links": null |
| }, |
| "BIBREF19": { |
| "ref_id": "b19", |
| "title": "Probabilistic Predictions of People Perusing: Evaluating Metrics of Language Model Performance for Psycholinguistic Modeling", |
| "authors": [ |
| { |
| "first": "Yiding", |
| "middle": [], |
| "last": "Hao", |
| "suffix": "" |
| }, |
| { |
| "first": "Simon", |
| "middle": [], |
| "last": "Mendelsohn", |
| "suffix": "" |
| }, |
| { |
| "first": "Rachel", |
| "middle": [], |
| "last": "Sterneck", |
| "suffix": "" |
| }, |
| { |
| "first": "Randi", |
| "middle": [], |
| "last": "Martinez", |
| "suffix": "" |
| }, |
| { |
| "first": "Robert", |
| "middle": [], |
| "last": "Frank", |
| "suffix": "" |
| } |
| ], |
| "year": 2020, |
| "venue": "Proceedings of the EMNLP Workshop on Cognitive Modeling and Computational Linguistics", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Yiding Hao, Simon Mendelsohn, Rachel Sterneck, Randi Martinez, and Robert Frank. 2020. Probabilis- tic Predictions of People Perusing: Evaluating Met- rics of Language Model Performance for Psycholin- guistic Modeling. In Proceedings of the EMNLP Workshop on Cognitive Modeling and Computational Linguistics.", |
| "links": null |
| }, |
| "BIBREF20": { |
| "ref_id": "b20", |
| "title": "The Copenhagen Corpus of Eye-Tracking Recordings from Natural Reading of Danish Texts", |
| "authors": [ |
| { |
| "first": "Nora", |
| "middle": [], |
| "last": "Hollenstein", |
| "suffix": "" |
| }, |
| { |
| "first": "Maria", |
| "middle": [], |
| "last": "Barrett", |
| "suffix": "" |
| }, |
| { |
| "first": "Marina", |
| "middle": [], |
| "last": "Bj\u00f6rnsd\u00f3ttir", |
| "suffix": "" |
| } |
| ], |
| "year": 2022, |
| "venue": "Proceedings of the 13th Conference on Language Resources and Evaluation (LREC 2022", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Nora Hollenstein, Maria Barrett, and Marina Bj\u00f6rnsd\u00f3t- tir. 2022. The Copenhagen Corpus of Eye-Tracking Recordings from Natural Reading of Danish Texts. In Proceedings of the 13th Conference on Language Resources and Evaluation (LREC 2022).", |
| "links": null |
| }, |
| "BIBREF21": { |
| "ref_id": "b21", |
| "title": "CMCL 2021 Shared Task on Eyetracking Prediction", |
| "authors": [ |
| { |
| "first": "Nora", |
| "middle": [], |
| "last": "Hollenstein", |
| "suffix": "" |
| }, |
| { |
| "first": "Emmanuele", |
| "middle": [], |
| "last": "Chersoni", |
| "suffix": "" |
| }, |
| { |
| "first": "Cassandra", |
| "middle": [ |
| "L" |
| ], |
| "last": "Jacobs", |
| "suffix": "" |
| }, |
| { |
| "first": "Yohei", |
| "middle": [], |
| "last": "Oseki", |
| "suffix": "" |
| }, |
| { |
| "first": "Laurent", |
| "middle": [], |
| "last": "Pr\u00e9vot", |
| "suffix": "" |
| }, |
| { |
| "first": "Enrico", |
| "middle": [], |
| "last": "Santus", |
| "suffix": "" |
| } |
| ], |
| "year": 2021, |
| "venue": "Proceedings of the Workshop on Cognitive Modeling and Computational Linguistics", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Nora Hollenstein, Emmanuele Chersoni, Cassandra L Jacobs, Yohei Oseki, Laurent Pr\u00e9vot, and Enrico Santus. 2021a. CMCL 2021 Shared Task on Eye- tracking Prediction. In Proceedings of the Workshop on Cognitive Modeling and Computational Linguis- tics.", |
| "links": null |
| }, |
| "BIBREF22": { |
| "ref_id": "b22", |
| "title": "Multilingual Language Models Predict Human Reading Behavior", |
| "authors": [ |
| { |
| "first": "Nora", |
| "middle": [], |
| "last": "Hollenstein", |
| "suffix": "" |
| }, |
| { |
| "first": "Federico", |
| "middle": [], |
| "last": "Pirovano", |
| "suffix": "" |
| }, |
| { |
| "first": "Ce", |
| "middle": [], |
| "last": "Zhang", |
| "suffix": "" |
| }, |
| { |
| "first": "Lena", |
| "middle": [], |
| "last": "J\u00e4ger", |
| "suffix": "" |
| }, |
| { |
| "first": "Lisa", |
| "middle": [], |
| "last": "Beinborn", |
| "suffix": "" |
| } |
| ], |
| "year": 2021, |
| "venue": "Proceedings of NAACL", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Nora Hollenstein, Federico Pirovano, Ce Zhang, Lena J\u00e4ger, and Lisa Beinborn. 2021b. Multilingual Lan- guage Models Predict Human Reading Behavior. In Proceedings of NAACL.", |
| "links": null |
| }, |
| "BIBREF23": { |
| "ref_id": "b23", |
| "title": "ZuCo, a Simultaneous EEG and Eye-tracking Resource for Natural Sentence Reading", |
| "authors": [ |
| { |
| "first": "Nora", |
| "middle": [], |
| "last": "Hollenstein", |
| "suffix": "" |
| }, |
| { |
| "first": "Jonathan", |
| "middle": [], |
| "last": "Rotsztejn", |
| "suffix": "" |
| }, |
| { |
| "first": "Marius", |
| "middle": [], |
| "last": "Troendle", |
| "suffix": "" |
| }, |
| { |
| "first": "Andreas", |
| "middle": [], |
| "last": "Pedroni", |
| "suffix": "" |
| }, |
| { |
| "first": "Ce", |
| "middle": [], |
| "last": "Zhang", |
| "suffix": "" |
| }, |
| { |
| "first": "Nicolas", |
| "middle": [], |
| "last": "Langer", |
| "suffix": "" |
| } |
| ], |
| "year": 2018, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Nora Hollenstein, Jonathan Rotsztejn, Marius Troen- dle, Andreas Pedroni, Ce Zhang, and Nicolas Langer. 2018. ZuCo, a Simultaneous EEG and Eye-tracking Resource for Natural Sentence Reading. Scientific Data.", |
| "links": null |
| }, |
| "BIBREF24": { |
| "ref_id": "b24", |
| "title": "ZuCo 2.0: A Dataset of Physiological Recordings During Natural Reading and Annotation", |
| "authors": [ |
| { |
| "first": "Nora", |
| "middle": [], |
| "last": "Hollenstein", |
| "suffix": "" |
| }, |
| { |
| "first": "Marius", |
| "middle": [], |
| "last": "Troendle", |
| "suffix": "" |
| }, |
| { |
| "first": "Ce", |
| "middle": [], |
| "last": "Zhang", |
| "suffix": "" |
| }, |
| { |
| "first": "Nicolas", |
| "middle": [], |
| "last": "Langer", |
| "suffix": "" |
| } |
| ], |
| "year": 2020, |
| "venue": "Proceedings of LREC", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Nora Hollenstein, Marius Troendle, Ce Zhang, and Nicolas Langer. 2020. ZuCo 2.0: A Dataset of Phys- iological Recordings During Natural Reading and Annotation. In Proceedings of LREC.", |
| "links": null |
| }, |
| "BIBREF25": { |
| "ref_id": "b25", |
| "title": "Integration and Prediction Difficulty in Hindi Sentence Comprehension: Evidence from an Eye-Tracking Corpus", |
| "authors": [ |
| { |
| "first": "Samar", |
| "middle": [], |
| "last": "Husain", |
| "suffix": "" |
| }, |
| { |
| "first": "Shravan", |
| "middle": [], |
| "last": "Vasishth", |
| "suffix": "" |
| }, |
| { |
| "first": "Narayanan", |
| "middle": [], |
| "last": "Srinivasan", |
| "suffix": "" |
| } |
| ], |
| "year": 2015, |
| "venue": "Journal of Eye Movement Research", |
| "volume": "8", |
| "issue": "2", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Samar Husain, Shravan Vasishth, and Narayanan Srini- vasan. 2015. Integration and Prediction Difficulty in Hindi Sentence Comprehension: Evidence from an Eye-Tracking Corpus. Journal of Eye Movement Research, 8(2).", |
| "links": null |
| }, |
| "BIBREF26": { |
| "ref_id": "b26", |
| "title": "NU HLT at CMCL 2022 Shared Task: Multilingual and Crosslingual Prediction of Human Reading Behavior in Universal Language Space", |
| "authors": [ |
| { |
| "first": "Joseph", |
| "middle": [], |
| "last": "Marvin Imperial", |
| "suffix": "" |
| } |
| ], |
| "year": 2022, |
| "venue": "Proceedings of the Workshop on Cognitive Modeling and Computational Linguistics", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Joseph Marvin Imperial. 2022. NU HLT at CMCL 2022 Shared Task: Multilingual and Crosslingual Prediction of Human Reading Behavior in Universal Language Space. In Proceedings of the Workshop on Cognitive Modeling and Computational Linguistics.", |
| "links": null |
| }, |
| "BIBREF27": { |
| "ref_id": "b27", |
| "title": "Potsdam Textbook Corpus (PoTeC): Eye Tracking Data from Experts and Non-experts Reading Scientific Texts", |
| "authors": [ |
| { |
| "first": "Lena", |
| "middle": [], |
| "last": "J\u00e4ger", |
| "suffix": "" |
| }, |
| { |
| "first": "Thomas", |
| "middle": [], |
| "last": "Kern", |
| "suffix": "" |
| }, |
| { |
| "first": "Patrick", |
| "middle": [], |
| "last": "Haller", |
| "suffix": "" |
| } |
| ], |
| "year": 2021, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": { |
| "DOI": [ |
| "10.17605/OSF.IO/DN5HP" |
| ] |
| }, |
| "num": null, |
| "urls": [], |
| "raw_text": "Lena J\u00e4ger, Thomas Kern, and Patrick Haller. 2021. Potsdam Textbook Corpus (PoTeC): Eye Track- ing Data from Experts and Non-experts Read- ing Scientific Texts. Available on OSF, DOI 10.17605/OSF.IO/DN5HP.", |
| "links": null |
| }, |
| "BIBREF28": { |
| "ref_id": "b28", |
| "title": "Tracking the Mind During Reading: The Influence of Past, Present, and Future Words on Fixation Durations", |
| "authors": [ |
| { |
| "first": "Reinhold", |
| "middle": [], |
| "last": "Kliegl", |
| "suffix": "" |
| }, |
| { |
| "first": "Antje", |
| "middle": [], |
| "last": "Nuthmann", |
| "suffix": "" |
| }, |
| { |
| "first": "Ralf", |
| "middle": [], |
| "last": "Engbert", |
| "suffix": "" |
| } |
| ], |
| "year": 2006, |
| "venue": "Journal of Experimental Psychology", |
| "volume": "135", |
| "issue": "1", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Reinhold Kliegl, Antje Nuthmann, and Ralf Engbert. 2006. Tracking the Mind During Reading: The In- fluence of Past, Present, and Future Words on Fixa- tion Durations. Journal of Experimental Psychology, 135(1):12.", |
| "links": null |
| }, |
| "BIBREF29": { |
| "ref_id": "b29", |
| "title": "Deepfix: A Fully Convolutional Neural Network for Predicting Human Eye Fixations", |
| "authors": [ |
| { |
| "first": "S", |
| "middle": [ |
| "S" |
| ], |
| "last": "Srinivas", |
| "suffix": "" |
| }, |
| { |
| "first": "Kumar", |
| "middle": [], |
| "last": "Kruthiventi", |
| "suffix": "" |
| }, |
| { |
| "first": "R Venkatesh", |
| "middle": [], |
| "last": "Ayush", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Babu", |
| "suffix": "" |
| } |
| ], |
| "year": 2017, |
| "venue": "IEEE Transactions on Image Processing", |
| "volume": "26", |
| "issue": "9", |
| "pages": "4446--4456", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Srinivas SS Kruthiventi, Kumar Ayush, and R Venkatesh Babu. 2017. Deepfix: A Fully Convolutional Neural Network for Predicting Human Eye Fixations. IEEE Transactions on Image Processing, 26(9):4446-4456.", |
| "links": null |
| }, |
| "BIBREF30": { |
| "ref_id": "b30", |
| "title": "Lower Perplexity Is Not Always Human-like", |
| "authors": [ |
| { |
| "first": "Tatsuki", |
| "middle": [], |
| "last": "Kuribayashi", |
| "suffix": "" |
| }, |
| { |
| "first": "Yohei", |
| "middle": [], |
| "last": "Oseki", |
| "suffix": "" |
| }, |
| { |
| "first": "Takumi", |
| "middle": [], |
| "last": "Ito", |
| "suffix": "" |
| }, |
| { |
| "first": "Ryo", |
| "middle": [], |
| "last": "Yoshida", |
| "suffix": "" |
| }, |
| { |
| "first": "Masayuki", |
| "middle": [], |
| "last": "Asahara", |
| "suffix": "" |
| }, |
| { |
| "first": "Kentaro", |
| "middle": [], |
| "last": "Inui", |
| "suffix": "" |
| } |
| ], |
| "year": 2021, |
| "venue": "Proceedings of ACL", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Tatsuki Kuribayashi, Yohei Oseki, Takumi Ito, Ryo Yoshida, Masayuki Asahara, and Kentaro Inui. 2021. Lower Perplexity Is Not Always Human-like. In Proceedings of ACL.", |
| "links": null |
| }, |
| "BIBREF31": { |
| "ref_id": "b31", |
| "title": "Crosslingual language model pretraining", |
| "authors": [ |
| { |
| "first": "Guillaume", |
| "middle": [], |
| "last": "Lample", |
| "suffix": "" |
| }, |
| { |
| "first": "Alexis", |
| "middle": [], |
| "last": "Conneau", |
| "suffix": "" |
| } |
| ], |
| "year": 2019, |
| "venue": "Advances in Neural Information Processing Systems", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Guillaume Lample and Alexis Conneau. 2019. Cross- lingual language model pretraining. In Advances in Neural Information Processing Systems.", |
| "links": null |
| }, |
| "BIBREF32": { |
| "ref_id": "b32", |
| "title": "Russian Sentence Corpus: Benchmark Measures of Eye Movements in Reading in Cyrillic", |
| "authors": [ |
| { |
| "first": "", |
| "middle": [], |
| "last": "Ak Laurinavichyute", |
| "suffix": "" |
| }, |
| { |
| "first": "A", |
| "middle": [], |
| "last": "Irina", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Sekerina", |
| "suffix": "" |
| }, |
| { |
| "first": "K", |
| "middle": [ |
| "A" |
| ], |
| "last": "Sv Alexeeva", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Bagdasaryan", |
| "suffix": "" |
| } |
| ], |
| "year": 2019, |
| "venue": "Behavior Research Methods", |
| "volume": "51", |
| "issue": "3", |
| "pages": "1161--1178", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "AK Laurinavichyute, Irina A Sekerina, SV Alexeeva, and KA Bagdasaryan. 2019. Russian Sentence Cor- pus: Benchmark Measures of Eye Movements in Reading in Cyrillic. Behavior Research Methods, 51(3):1161-1178.", |
| "links": null |
| }, |
| "BIBREF33": { |
| "ref_id": "b33", |
| "title": "Expectation-based Syntactic Comprehension", |
| "authors": [ |
| { |
| "first": "Roger", |
| "middle": [], |
| "last": "Levy", |
| "suffix": "" |
| } |
| ], |
| "year": 2008, |
| "venue": "Cognition", |
| "volume": "106", |
| "issue": "3", |
| "pages": "1126--1177", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Roger Levy. 2008. Expectation-based Syntactic Com- prehension. Cognition, 106(3):1126-1177.", |
| "links": null |
| }, |
| "BIBREF34": { |
| "ref_id": "b34", |
| "title": "Universality in Eye Movements and Reading: A Trilingual Investigation. Cognition", |
| "authors": [ |
| { |
| "first": "Denis", |
| "middle": [], |
| "last": "Simon P Liversedge", |
| "suffix": "" |
| }, |
| { |
| "first": "Xin", |
| "middle": [], |
| "last": "Drieghe", |
| "suffix": "" |
| }, |
| { |
| "first": "Guoli", |
| "middle": [], |
| "last": "Li", |
| "suffix": "" |
| }, |
| { |
| "first": "Xuejun", |
| "middle": [], |
| "last": "Yan", |
| "suffix": "" |
| }, |
| { |
| "first": "Jukka", |
| "middle": [], |
| "last": "Bai", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Hy\u00f6n\u00e4", |
| "suffix": "" |
| } |
| ], |
| "year": 2016, |
| "venue": "", |
| "volume": "147", |
| "issue": "", |
| "pages": "1--20", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Simon P Liversedge, Denis Drieghe, Xin Li, Guoli Yan, Xuejun Bai, and Jukka Hy\u00f6n\u00e4. 2016. Universality in Eye Movements and Reading: A Trilingual Investi- gation. Cognition, 147:1-20.", |
| "links": null |
| }, |
| "BIBREF35": { |
| "ref_id": "b35", |
| "title": "The Provo Corpus: A Large Eye-tracking Corpus with Predictability Norms", |
| "authors": [ |
| { |
| "first": "G", |
| "middle": [], |
| "last": "Steven", |
| "suffix": "" |
| }, |
| { |
| "first": "Kiel", |
| "middle": [], |
| "last": "Luke", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Christianson", |
| "suffix": "" |
| } |
| ], |
| "year": 2017, |
| "venue": "Behavior Research Methods", |
| "volume": "", |
| "issue": "", |
| "pages": "1--8", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Steven G Luke and Kiel Christianson. 2017. The Provo Corpus: A Large Eye-tracking Corpus with Predictability Norms. Behavior Research Methods, pages 1-8.", |
| "links": null |
| }, |
| "BIBREF36": { |
| "ref_id": "b36", |
| "title": "With Blinkers On: Robust Prediction of Eye Movements across Readers", |
| "authors": [ |
| { |
| "first": "Franz", |
| "middle": [], |
| "last": "Matthies", |
| "suffix": "" |
| }, |
| { |
| "first": "Anders", |
| "middle": [], |
| "last": "S\u00f8gaard", |
| "suffix": "" |
| } |
| ], |
| "year": 2013, |
| "venue": "Proceedings of EMNLP", |
| "volume": "", |
| "issue": "", |
| "pages": "803--807", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Franz Matthies and Anders S\u00f8gaard. 2013. With Blink- ers On: Robust Prediction of Eye Movements across Readers. Proceedings of EMNLP, pages 803-807.", |
| "links": null |
| }, |
| "BIBREF37": { |
| "ref_id": "b37", |
| "title": "Relation Classification with Cognitive Attention Supervision", |
| "authors": [ |
| { |
| "first": "Erik", |
| "middle": [], |
| "last": "Mcguire", |
| "suffix": "" |
| }, |
| { |
| "first": "Noriko", |
| "middle": [], |
| "last": "Tomuro", |
| "suffix": "" |
| } |
| ], |
| "year": 2021, |
| "venue": "Proceedings of the NAACL Workshop on Cognitive Modeling and Computational Linguistics", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Erik McGuire and Noriko Tomuro. 2021. Relation Clas- sification with Cognitive Attention Supervision. In Proceedings of the NAACL Workshop on Cognitive Modeling and Computational Linguistics.", |
| "links": null |
| }, |
| "BIBREF38": { |
| "ref_id": "b38", |
| "title": "Human Sentence Processing: Recurrence or Attention?", |
| "authors": [ |
| { |
| "first": "Danny", |
| "middle": [], |
| "last": "Merkx", |
| "suffix": "" |
| }, |
| { |
| "first": "L", |
| "middle": [], |
| "last": "Stefan", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Frank", |
| "suffix": "" |
| } |
| ], |
| "year": null, |
| "venue": "Proceedings of the NAACL Workshop on Cognitive Modeling and Computational Linguistics", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Danny Merkx and Stefan L Frank. Human Sentence Processing: Recurrence or Attention? In Proceed- ings of the NAACL Workshop on Cognitive Modeling and Computational Linguistics.", |
| "links": null |
| }, |
| "BIBREF39": { |
| "ref_id": "b39", |
| "title": "Syntactic and Semantic Factors in Processing Difficulty: An Integrated Measure", |
| "authors": [ |
| { |
| "first": "Jeff", |
| "middle": [], |
| "last": "Mitchell", |
| "suffix": "" |
| }, |
| { |
| "first": "Mirella", |
| "middle": [], |
| "last": "Lapata", |
| "suffix": "" |
| }, |
| { |
| "first": "Vera", |
| "middle": [], |
| "last": "Demberg", |
| "suffix": "" |
| }, |
| { |
| "first": "Frank", |
| "middle": [], |
| "last": "Keller", |
| "suffix": "" |
| } |
| ], |
| "year": 2010, |
| "venue": "Proceedings of ACL", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Jeff Mitchell, Mirella Lapata, Vera Demberg, and Frank Keller. 2010. Syntactic and Semantic Factors in Pro- cessing Difficulty: An Integrated Measure. In Pro- ceedings of ACL.", |
| "links": null |
| }, |
| "BIBREF40": { |
| "ref_id": "b40", |
| "title": "The Beijing Sentence Corpus: A Chinese Sentence Corpus with Eye Movement Data and Predictability Norms", |
| "authors": [ |
| { |
| "first": "Jinger", |
| "middle": [], |
| "last": "Pan", |
| "suffix": "" |
| }, |
| { |
| "first": "Ming", |
| "middle": [], |
| "last": "Yan", |
| "suffix": "" |
| }, |
| { |
| "first": "M", |
| "middle": [], |
| "last": "Eike", |
| "suffix": "" |
| }, |
| { |
| "first": "Hua", |
| "middle": [], |
| "last": "Richter", |
| "suffix": "" |
| }, |
| { |
| "first": "Reinhold", |
| "middle": [], |
| "last": "Shu", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Kliegl", |
| "suffix": "" |
| } |
| ], |
| "year": 2021, |
| "venue": "Behavior Research Methods", |
| "volume": "", |
| "issue": "", |
| "pages": "1--12", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Jinger Pan, Ming Yan, Eike M Richter, Hua Shu, and Reinhold Kliegl. 2021. The Beijing Sentence Cor- pus: A Chinese Sentence Corpus with Eye Movement Data and Predictability Norms. Behavior Research Methods, pages 1-12.", |
| "links": null |
| }, |
| "BIBREF41": { |
| "ref_id": "b41", |
| "title": "AdapterHub: A Framework for Adapting Transformers", |
| "authors": [ |
| { |
| "first": "Jonas", |
| "middle": [], |
| "last": "Pfeiffer", |
| "suffix": "" |
| }, |
| { |
| "first": "Andreas", |
| "middle": [], |
| "last": "R\u00fcckl\u00e9", |
| "suffix": "" |
| }, |
| { |
| "first": "Clifton", |
| "middle": [], |
| "last": "Poth", |
| "suffix": "" |
| }, |
| { |
| "first": "Aishwarya", |
| "middle": [], |
| "last": "Kamath", |
| "suffix": "" |
| }, |
| { |
| "first": "Ivan", |
| "middle": [], |
| "last": "Vuli\u0107", |
| "suffix": "" |
| }, |
| { |
| "first": "Sebastian", |
| "middle": [], |
| "last": "Ruder", |
| "suffix": "" |
| }, |
| { |
| "first": "Kyunghyun", |
| "middle": [], |
| "last": "Cho", |
| "suffix": "" |
| }, |
| { |
| "first": "Iryna", |
| "middle": [], |
| "last": "Gurevych", |
| "suffix": "" |
| } |
| ], |
| "year": 2020, |
| "venue": "Proceedings of EMNLP: System Demonstrations", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Jonas Pfeiffer, Andreas R\u00fcckl\u00e9, Clifton Poth, Aishwarya Kamath, Ivan Vuli\u0107, Sebastian Ruder, Kyunghyun Cho, and Iryna Gurevych. 2020. AdapterHub: A Framework for Adapting Transformers. In Proceed- ings of EMNLP: System Demonstrations.", |
| "links": null |
| }, |
| "BIBREF42": { |
| "ref_id": "b42", |
| "title": "Language models Are Unsupervised Multitask Learners", |
| "authors": [ |
| { |
| "first": "Alec", |
| "middle": [], |
| "last": "Radford", |
| "suffix": "" |
| }, |
| { |
| "first": "Jeffrey", |
| "middle": [], |
| "last": "Wu", |
| "suffix": "" |
| }, |
| { |
| "first": "Rewon", |
| "middle": [], |
| "last": "Child", |
| "suffix": "" |
| }, |
| { |
| "first": "David", |
| "middle": [], |
| "last": "Luan", |
| "suffix": "" |
| }, |
| { |
| "first": "Dario", |
| "middle": [], |
| "last": "Amodei", |
| "suffix": "" |
| }, |
| { |
| "first": "Ilya", |
| "middle": [], |
| "last": "Sutskever", |
| "suffix": "" |
| } |
| ], |
| "year": 2019, |
| "venue": "Ope-nAI blog", |
| "volume": "1", |
| "issue": "8", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Alec Radford, Jeffrey Wu, Rewon Child, David Luan, Dario Amodei, Ilya Sutskever, et al. 2019. Language models Are Unsupervised Multitask Learners. Ope- nAI blog, 1(8):9.", |
| "links": null |
| }, |
| "BIBREF43": { |
| "ref_id": "b43", |
| "title": "Eye Movement Control in Reading and Visual Search: Effects of Word Frequency", |
| "authors": [ |
| { |
| "first": "Keith", |
| "middle": [], |
| "last": "Rayner", |
| "suffix": "" |
| }, |
| { |
| "first": "Gary", |
| "middle": [ |
| "E" |
| ], |
| "last": "Raney", |
| "suffix": "" |
| } |
| ], |
| "year": 1996, |
| "venue": "Psychonomic Bulletin & Review", |
| "volume": "3", |
| "issue": "2", |
| "pages": "245--248", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Keith Rayner and Gary E Raney. 1996. Eye Movement Control in Reading and Visual Search: Effects of Word Frequency. Psychonomic Bulletin & Review, 3(2):245-248.", |
| "links": null |
| }, |
| "BIBREF44": { |
| "ref_id": "b44", |
| "title": "Toward a Model of Eye Movement Control in Reading. Psychological Review", |
| "authors": [ |
| { |
| "first": "Alexander", |
| "middle": [], |
| "last": "Erik D Reichle", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Pollatsek", |
| "suffix": "" |
| }, |
| { |
| "first": "L", |
| "middle": [], |
| "last": "Donald", |
| "suffix": "" |
| }, |
| { |
| "first": "Keith", |
| "middle": [], |
| "last": "Fisher", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Rayner", |
| "suffix": "" |
| } |
| ], |
| "year": 1998, |
| "venue": "", |
| "volume": "105", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Erik D Reichle, Alexander Pollatsek, Donald L Fisher, and Keith Rayner. 1998. Toward a Model of Eye Movement Control in Reading. Psychological Re- view, 105(1):125.", |
| "links": null |
| }, |
| "BIBREF45": { |
| "ref_id": "b45", |
| "title": "HkAmsters at CMCL 2022 Shared Task: Predicting Eye-Tracking Data from a Gradient Boosting Framework with Linguistic Features", |
| "authors": [ |
| { |
| "first": "L", |
| "middle": [], |
| "last": "Salicchi", |
| "suffix": "" |
| }, |
| { |
| "first": "R", |
| "middle": [], |
| "last": "Xiang", |
| "suffix": "" |
| }, |
| { |
| "first": "Y", |
| "middle": [], |
| "last": "Hsu", |
| "suffix": "" |
| } |
| ], |
| "year": 2022, |
| "venue": "Proceedings of the Workshop on Cognitive Modeling and Computational Linguistics", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "L. Salicchi, R. Xiang, and Y. Hsu. 2022. HkAmsters at CMCL 2022 Shared Task: Predicting Eye-Tracking Data from a Gradient Boosting Framework with Lin- guistic Features. In Proceedings of the Workshop on Cognitive Modeling and Computational Linguistics.", |
| "links": null |
| }, |
| "BIBREF46": { |
| "ref_id": "b46", |
| "title": "Looking for a Role for Word Embeddings in Eye-Tracking Features Prediction: Does Semantic Similarity Help?", |
| "authors": [ |
| { |
| "first": "Lavinia", |
| "middle": [], |
| "last": "Salicchi", |
| "suffix": "" |
| }, |
| { |
| "first": "Alessandro", |
| "middle": [], |
| "last": "Lenci", |
| "suffix": "" |
| }, |
| { |
| "first": "Emmanuele", |
| "middle": [], |
| "last": "Chersoni", |
| "suffix": "" |
| } |
| ], |
| "year": 2021, |
| "venue": "Proceedings of IWCS", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Lavinia Salicchi, Alessandro Lenci, and Emmanuele Chersoni. 2021. Looking for a Role for Word Em- beddings in Eye-Tracking Features Prediction: Does Semantic Similarity Help? In Proceedings of IWCS.", |
| "links": null |
| }, |
| "BIBREF47": { |
| "ref_id": "b47", |
| "title": "That Looks Hard: Characterizing Linguistic Complexity in Humans and Language Models", |
| "authors": [ |
| { |
| "first": "Gabriele", |
| "middle": [], |
| "last": "Sarti", |
| "suffix": "" |
| }, |
| { |
| "first": "Dominique", |
| "middle": [], |
| "last": "Brunato", |
| "suffix": "" |
| }, |
| { |
| "first": "Felice", |
| "middle": [], |
| "last": "Dell'orletta", |
| "suffix": "" |
| } |
| ], |
| "year": 2021, |
| "venue": "Proceedings of the NAACL Workshop on Cognitive Modeling and Computational Linguistics", |
| "volume": "", |
| "issue": "", |
| "pages": "48--60", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Gabriele Sarti, Dominique Brunato, and Felice Dell'Orletta. 2021. That Looks Hard: Characteriz- ing Linguistic Complexity in Humans and Language Models. In Proceedings of the NAACL Workshop on Cognitive Modeling and Computational Linguistics, pages 48-60.", |
| "links": null |
| }, |
| "BIBREF48": { |
| "ref_id": "b48", |
| "title": "Eye Tracking Assisted Extraction of Attentionally Important Objects from Videos", |
| "authors": [ |
| { |
| "first": "Thuyen", |
| "middle": [], |
| "last": "Karthikeyan Shanmuga Vadivel", |
| "suffix": "" |
| }, |
| { |
| "first": "Miguel", |
| "middle": [], |
| "last": "Ngo", |
| "suffix": "" |
| }, |
| { |
| "first": "B", |
| "middle": [ |
| "S" |
| ], |
| "last": "Eckstein", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Manjunath", |
| "suffix": "" |
| } |
| ], |
| "year": 2015, |
| "venue": "Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition", |
| "volume": "", |
| "issue": "", |
| "pages": "3241--3250", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Karthikeyan Shanmuga Vadivel, Thuyen Ngo, Miguel Eckstein, and BS Manjunath. 2015. Eye Tracking As- sisted Extraction of Attentionally Important Objects from Videos. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 3241-3250.", |
| "links": null |
| }, |
| "BIBREF49": { |
| "ref_id": "b49", |
| "title": "Expanding Horizons of Cross-linguistic Research on Reading: The Multilingual Eye-Movement Corpus (MECO)", |
| "authors": [ |
| { |
| "first": "Noam", |
| "middle": [], |
| "last": "Siegelman", |
| "suffix": "" |
| }, |
| { |
| "first": "Sascha", |
| "middle": [], |
| "last": "Schroeder", |
| "suffix": "" |
| }, |
| { |
| "first": "V", |
| "middle": [], |
| "last": "Kuperman", |
| "suffix": "" |
| } |
| ], |
| "year": 2022, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Noam Siegelman, Sascha Schroeder, and V Kuperman. 2022. Expanding Horizons of Cross-linguistic Re- search on Reading: The Multilingual Eye-Movement Corpus (MECO). Behavior Research Methods.", |
| "links": null |
| }, |
| "BIBREF50": { |
| "ref_id": "b50", |
| "title": "Quantifying Sentence Complexity Based on Eye-tracking Measures", |
| "authors": [ |
| { |
| "first": "Abhinav Deep", |
| "middle": [], |
| "last": "Singh", |
| "suffix": "" |
| }, |
| { |
| "first": "Poojan", |
| "middle": [], |
| "last": "Mehta", |
| "suffix": "" |
| }, |
| { |
| "first": "Samar", |
| "middle": [], |
| "last": "Husain", |
| "suffix": "" |
| }, |
| { |
| "first": "Rajkumar", |
| "middle": [], |
| "last": "Rajakrishnan", |
| "suffix": "" |
| } |
| ], |
| "year": 2016, |
| "venue": "Proceedings of the COLING Workshop on Computational Linguistics for Linguistic Complexity", |
| "volume": "", |
| "issue": "", |
| "pages": "202--212", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Abhinav Deep Singh, Poojan Mehta, Samar Husain, and Rajkumar Rajakrishnan. 2016. Quantifying Sentence Complexity Based on Eye-tracking Measures. In Proceedings of the COLING Workshop on Computa- tional Linguistics for Linguistic Complexity, pages 202-212.", |
| "links": null |
| }, |
| "BIBREF51": { |
| "ref_id": "b51", |
| "title": "Poirot at CMCL 2022 Shared Task: Zero Shot Crosslingual Eye-Tracking Data Prediction using Multilingual Transformer Models", |
| "authors": [ |
| { |
| "first": "Harshvardhan", |
| "middle": [], |
| "last": "Srivastava", |
| "suffix": "" |
| } |
| ], |
| "year": 2022, |
| "venue": "Proceedings of the Workshop on Cognitive Modeling and Computational Linguistics", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Harshvardhan Srivastava. 2022. Poirot at CMCL 2022 Shared Task: Zero Shot Crosslingual Eye-Tracking Data Prediction using Multilingual Transformer Mod- els. In Proceedings of the Workshop on Cognitive Modeling and Computational Linguistics.", |
| "links": null |
| }, |
| "BIBREF52": { |
| "ref_id": "b52", |
| "title": "Team DMG at CMCL 2022 Shared Task: Transformer Adapters for the Multi-and Cross-Lingual Prediction of Human Reading Behavior", |
| "authors": [], |
| "year": null, |
| "venue": "Proceedings of the ACL Workshop on Cognitive Modeling and Computational Linguistics", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Ece Takmaz. 2022. Team DMG at CMCL 2022 Shared Task: Transformer Adapters for the Multi-and Cross- Lingual Prediction of Human Reading Behavior. In Proceedings of the ACL Workshop on Cognitive Mod- eling and Computational Linguistics.", |
| "links": null |
| }, |
| "BIBREF53": { |
| "ref_id": "b53", |
| "title": "Eye-tracking Data for Weakly Supervised Object Detection", |
| "authors": [ |
| { |
| "first": "Ching-Hsi", |
| "middle": [], |
| "last": "Tseng", |
| "suffix": "" |
| }, |
| { |
| "first": "Yen", |
| "middle": [], |
| "last": "Hsu", |
| "suffix": "" |
| }, |
| { |
| "first": "Shyan-Ming", |
| "middle": [], |
| "last": "Yuan", |
| "suffix": "" |
| } |
| ], |
| "year": 2020, |
| "venue": "2020 IEEE Eurasia Conference on IOT, Communication and Engineering (ECICE)", |
| "volume": "", |
| "issue": "", |
| "pages": "223--225", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Ching-Hsi Tseng, Yen Hsu, and Shyan-Ming Yuan. 2020. Eye-tracking Data for Weakly Supervised Ob- ject Detection. In 2020 IEEE Eurasia Conference on IOT, Communication and Engineering (ECICE), pages 223-225.", |
| "links": null |
| }, |
| "BIBREF54": { |
| "ref_id": "b54", |
| "title": "On the Predictive Power of Neural Language Models for Human Real-Time Comprehension Behavior", |
| "authors": [ |
| { |
| "first": "Ethan", |
| "middle": [ |
| "Gotlieb" |
| ], |
| "last": "Wilcox", |
| "suffix": "" |
| }, |
| { |
| "first": "Jon", |
| "middle": [], |
| "last": "Gauthier", |
| "suffix": "" |
| }, |
| { |
| "first": "Jennifer", |
| "middle": [], |
| "last": "Hu", |
| "suffix": "" |
| }, |
| { |
| "first": "Peng", |
| "middle": [], |
| "last": "Qian", |
| "suffix": "" |
| }, |
| { |
| "first": "Roger", |
| "middle": [], |
| "last": "Levy", |
| "suffix": "" |
| } |
| ], |
| "year": 2020, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": { |
| "arXiv": [ |
| "arXiv:2006.01912" |
| ] |
| }, |
| "num": null, |
| "urls": [], |
| "raw_text": "Ethan Gotlieb Wilcox, Jon Gauthier, Jennifer Hu, Peng Qian, and Roger Levy. 2020. On the Predictive Power of Neural Language Models for Human Real- Time Comprehension Behavior. arXiv preprint arXiv:2006.01912.", |
| "links": null |
| }, |
| "BIBREF55": { |
| "ref_id": "b55", |
| "title": "Eye Movements, Word Familiarity, and Vocabulary Acquisition", |
| "authors": [ |
| { |
| "first": "Rihana", |
| "middle": [], |
| "last": "Williams", |
| "suffix": "" |
| }, |
| { |
| "first": "Robin", |
| "middle": [], |
| "last": "Morris", |
| "suffix": "" |
| } |
| ], |
| "year": 2004, |
| "venue": "European Journal of Cognitive Psychology", |
| "volume": "16", |
| "issue": "1-2", |
| "pages": "312--339", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Rihana Williams and Robin Morris. 2004. Eye Move- ments, Word Familiarity, and Vocabulary Acquisi- tion. European Journal of Cognitive Psychology, 16(1-2):312-339.", |
| "links": null |
| }, |
| "BIBREF56": { |
| "ref_id": "b56", |
| "title": "KonTra at CMCL 2021 Shared Task: Predicting Eye Movements by Combining BERT with Surface, Linguistic and Behavioral Information", |
| "authors": [ |
| { |
| "first": "Qi", |
| "middle": [], |
| "last": "Yu", |
| "suffix": "" |
| }, |
| { |
| "first": "Aikaterini-Lida", |
| "middle": [], |
| "last": "Kalouli", |
| "suffix": "" |
| }, |
| { |
| "first": "Diego", |
| "middle": [], |
| "last": "Frassinelli", |
| "suffix": "" |
| } |
| ], |
| "year": 2021, |
| "venue": "Proceedings of the NAACL Workshop on Cognitive Modeling and Computational Linguistics", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Qi Yu, Aikaterini-Lida Kalouli, and Diego Frassinelli. 2021. KonTra at CMCL 2021 Shared Task: Pre- dicting Eye Movements by Combining BERT with Surface, Linguistic and Behavioral Information. In Proceedings of the NAACL Workshop on Cognitive Modeling and Computational Linguistics.", |
| "links": null |
| } |
| }, |
| "ref_entries": { |
| "FIGREF0": { |
| "num": null, |
| "type_str": "figure", |
| "text": "An example sentence from the Russian Sentence Corpus (Laurinavichyute et al., 2019) averaged across all readers. A wider diameter of the markers represents a higher standard deviation.", |
| "uris": null |
| }, |
| "TABREF0": { |
| "type_str": "table", |
| "text": "Lang. Sents. Tokens Subjects Reference", |
| "html": null, |
| "content": "<table><tr><td>BSC</td><td>ZH</td><td>150</td><td>1685</td><td>60 Pan et al. (2021)</td></tr><tr><td>PAHEC</td><td>HI</td><td>153</td><td>2596</td><td>30 Husain et al. (2015)</td></tr><tr><td>RSC</td><td>RU</td><td>144</td><td>1417</td><td>102 Laurinavichyute et al. (2019)</td></tr><tr><td>Provo</td><td>EN</td><td>189</td><td>2659</td><td>84 Luke and Christianson (2017)</td></tr><tr><td>ZuCo 1.0</td><td>EN</td><td>300</td><td>6588</td><td>12 Hollenstein et al. (2018)</td></tr><tr><td>ZuCo 2.0</td><td>EN</td><td>349</td><td>6828</td><td>18 Hollenstein et al. (2020)</td></tr><tr><td>GECO-NL</td><td>NL</td><td>800</td><td>9218</td><td>18 Cop et al. (2017)</td></tr><tr><td>PoTeC</td><td>DE</td><td>101</td><td>1895</td><td>75 J\u00e4ger et al. (2021)</td></tr><tr><td>CopCo (Subtask 2 only)</td><td>DK</td><td>402</td><td>6768</td><td>5 Hollenstein et al. (2022)</td></tr></table>", |
| "num": null |
| }, |
| "TABREF1": { |
| "type_str": "table", |
| "text": "Datasets used in the shared task. Note that the number of sentences and tokens refers to the text materials we have selected and not necessarily to the complete original datasets.", |
| "html": null, |
| "content": "<table><tr><td>Feature</td><td>min</td><td>max</td><td>mean (std)</td></tr><tr><td>FFDAVG</td><td colspan=\"2\">0.0 56.74</td><td>13.02 (7.34)</td></tr><tr><td>FFDSTD</td><td colspan=\"2\">0.0 58.54</td><td>4.47 (3.55)</td></tr><tr><td>TRTAVG</td><td colspan=\"3\">0.0 100.0 18.87 (11.57)</td></tr><tr><td>TRTSTD</td><td colspan=\"2\">0.0 100.0</td><td>9.86 (8.01)</td></tr></table>", |
| "num": null |
| }, |
| "TABREF2": { |
| "type_str": "table", |
| "text": "Minimum, maximum, mean and standard deviation of the scaled feature values in both training and test data of Subtask 1, after averaging across readers.", |
| "html": null, |
| "content": "<table/>", |
| "num": null |
| }, |
| "TABREF3": { |
| "type_str": "table", |
| "text": "The", |
| "html": null, |
| "content": "<table><tr><td/><td/><td>train</td><td/><td/></tr><tr><td>100</td><td/><td/><td/><td/></tr><tr><td>80</td><td/><td/><td/><td/></tr><tr><td>60</td><td/><td/><td/><td/></tr><tr><td>40</td><td/><td/><td/><td/></tr><tr><td>20</td><td/><td/><td/><td/></tr><tr><td>0</td><td>11.05</td><td>3.57</td><td>15.11</td><td>7.8</td></tr><tr><td/><td>FFDAvg</td><td>FFDStd</td><td>TRTAvg</td><td>TRTStd</td></tr><tr><td/><td/><td colspan=\"2\">test subtask 1</td><td/></tr><tr><td>100</td><td/><td/><td/><td/></tr><tr><td>80</td><td/><td/><td/><td/></tr><tr><td>60</td><td/><td/><td/><td/></tr><tr><td>40</td><td/><td/><td/><td/></tr><tr><td>20</td><td/><td/><td/><td/></tr><tr><td>0</td><td>11.25</td><td>3.63</td><td>15.4</td><td>7.85</td></tr><tr><td/><td>FFDAvg</td><td>FFDStd</td><td>TRTAvg</td><td>TRTStd</td></tr><tr><td/><td/><td colspan=\"2\">test subtask 2</td><td/></tr><tr><td>100</td><td/><td/><td/><td/></tr><tr><td>80</td><td/><td/><td/><td>4 Data</td></tr><tr><td>60</td><td/><td/><td/><td>4.1 Subtask 1</td></tr><tr><td>40</td><td/><td/><td/><td/></tr><tr><td>20</td><td/><td/><td/><td/></tr><tr><td>0</td><td>11.36</td><td>2.13</td><td>14.36</td><td>3.69</td></tr><tr><td/><td>FFDAvg</td><td>FFDStd</td><td>TRTAvg</td><td>TRTStd</td></tr><tr><td colspan=\"5\">Figure 2: Boxplots showing the feature value distri-butions of the training data and the test sets of both subtasks. Below each box is the median value of each feature.</td></tr><tr><td/><td/><td/><td/><td>These datasets</td></tr><tr><td/><td/><td/><td/><td>cover a diverse range of text domains, including</td></tr><tr><td/><td/><td/><td/><td>news articles, novels, Wikipedia sentences, scien-</td></tr><tr><td/><td/><td/><td/><td>tific textbook passages, etc. The details are pre-</td></tr><tr><td/><td/><td/><td/><td>sented in Table 1.</td></tr><tr><td/><td/><td/><td/><td>The training data contains 1703 sentences, the</td></tr><tr><td/><td/><td/><td/><td>development set contains 104 sentences, and the</td></tr><tr><td/><td/><td/><td/><td>test set 324 sentences.</td></tr><tr><td/><td/><td/><td/><td>4.2 Subtask 2</td></tr><tr><td/><td/><td/><td/><td>As described, the training and development data</td></tr><tr><td/><td/><td/><td/><td>are identical to Subtask 1, but the test data contains</td></tr><tr><td/><td/><td/><td/><td>eye-tracking data from a new language, namely</td></tr><tr><td/><td/><td/><td/><td>Danish. The Danish eye-tracking data contains 402</td></tr></table>", |
| "num": null |
| }, |
| "TABREF5": { |
| "type_str": "table", |
| "text": "Overall results of Subtask 1 showing the best submission per team and the mean baselines, including the overall MAE and R 2 scores, as well as the individual MAE scores for each feature. The teams are ranked by the MAE averaged across all five eye-tracking features (third column).", |
| "html": null, |
| "content": "<table><tr><td colspan=\"2\">Rank Team Name</td><td colspan=\"5\">MAE FFDAVG FFDSTD TRTAVG TRTSTD</td><td>R 2 Reference</td></tr><tr><td>1 2</td><td>Poirot DMG</td><td>4.23 4.97</td><td>5.60 6.90</td><td>5.65 5.77</td><td>2.95 5.45</td><td colspan=\"2\">2.73 -0.26 Srivastava (2022) 1.73 -0.57 Takmaz (2022)</td></tr><tr><td>-</td><td>Mean baseline</td><td>5.73</td><td>8.82</td><td>5.89</td><td>5.69</td><td>2.54</td><td>0.00 -</td></tr><tr><td>3</td><td>NU HLT</td><td>7.09</td><td>14.65</td><td>4.04</td><td>7.53</td><td colspan=\"2\">2.12 -1.83 Imperial (2022)</td></tr></table>", |
| "num": null |
| }, |
| "TABREF6": { |
| "type_str": "table", |
| "text": "Overall results of Subtask 2 showing the best submission per team and the mean baseline, including the overall MAE and R 2 scores, as well as the individual MAE scores for each feature. The teams are ranked by the MAE averaged across all five eye-tracking features (third column).", |
| "html": null, |
| "content": "<table/>", |
| "num": null |
| } |
| } |
| } |
| } |
