FindingsPub Date : 2024-02-23DOI: 10.48550/arXiv.2402.15120
Hyunjae Kim, Seunghyun Yoon, Trung Bui, Handong Zhao, Q. Tran, Franck Dernoncourt, Jaewoo Kang
{"title":"Fine-tuning CLIP Text Encoders with Two-step Paraphrasing","authors":"Hyunjae Kim, Seunghyun Yoon, Trung Bui, Handong Zhao, Q. Tran, Franck Dernoncourt, Jaewoo Kang","doi":"10.48550/arXiv.2402.15120","DOIUrl":"https://doi.org/10.48550/arXiv.2402.15120","url":null,"abstract":"Contrastive language-image pre-training (CLIP) models have demonstrated considerable success across various vision-language tasks, such as text-to-image retrieval, where the model is required to effectively process natural language input to produce an accurate visual output. However, current models still face limitations in dealing with linguistic variations in input queries, such as paraphrases, making it challenging to handle a broad range of user queries in real-world applications. In this study, we introduce a straightforward fine-tuning approach to enhance the representations of CLIP models for paraphrases. Our approach involves a two-step paraphrase generation process, where we automatically create two categories of paraphrases from web-scale image captions by leveraging large language models. Subsequently, we fine-tune the CLIP text encoder using these generated paraphrases while freezing the image encoder. Our resulting model, which we call ParaCLIP, exhibits significant improvements over baseline CLIP models across various tasks, including paraphrased retrieval (with rank similarity scores improved by up to 7.6% and 9.6%), Visual Genome Relation and Attribution, as well as seven semantic textual similarity tasks.","PeriodicalId":508951,"journal":{"name":"Findings","volume":"32 7","pages":"2175-2184"},"PeriodicalIF":0.0,"publicationDate":"2024-02-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140437106","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
FindingsPub Date : 2024-02-22DOI: 10.32866/001c.93917
Meredith Alousi-Jones, A. El-geneidy
{"title":"If Transit is Free will Older Adults use it More? A Longitudinal Analysis","authors":"Meredith Alousi-Jones, A. El-geneidy","doi":"10.32866/001c.93917","DOIUrl":"https://doi.org/10.32866/001c.93917","url":null,"abstract":"Free public transit for older adults is a policy often advocated for to increase ridership and reduce their dependence on automobiles. In July 2023 public transit became free on the island of Montréal for older adults (65+). Drawing from longitudinal repeated survey data collected in February 2023 and September 2023 and in-depth interviews collected in November 2023, we study the impacts of free public transit fare policy on older adults’ (65+) frequency of public transit use and satisfaction with service. We did not observe any change in public transit use among those who anticipated increased use in the before period, nor among those who anticipated no change in their use after the implementation of the free fare policy. We observed a positive change in the perception of the convenience of public transit after the implementation of the free fare policy for older adults.","PeriodicalId":508951,"journal":{"name":"Findings","volume":"23 8","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-02-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140441086","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
FindingsPub Date : 2024-02-20DOI: 10.32866/001c.93913
Mahtot Gebresselassie, Seyedmohsen Alavi, Andy Hong
{"title":"How and Where Did Older People Travel Before and After COVID? Insights from Washington, DC’s Smart Card Data","authors":"Mahtot Gebresselassie, Seyedmohsen Alavi, Andy Hong","doi":"10.32866/001c.93913","DOIUrl":"https://doi.org/10.32866/001c.93913","url":null,"abstract":"Using smart card data of subway trips, this paper analyzed travel behaviors of older adults in Washington, DC in three phases of COVID-19 (Pre: 2018-2019, Early: 2020, Late: 2021-2022). The findings show that the impact of COVID-19 on average daily travel patterns was more pronounced on weekday travels, compared to weekend trips. In addition, compared to the general population, older adults’ subway usage showed a slower recovery to normal patterns in both usage levels and trip destinations. The results reveal important insights for transportation planners and transit authorities about older adults’ travel patterns during normal times and unusual events.","PeriodicalId":508951,"journal":{"name":"Findings","volume":"1 6","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139958331","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
FindingsPub Date : 2024-02-20DOI: 10.48550/arXiv.2402.12880
Nadine Probol, Margot Mieskes
{"title":"Autism Detection in Speech – A Survey","authors":"Nadine Probol, Margot Mieskes","doi":"10.48550/arXiv.2402.12880","DOIUrl":"https://doi.org/10.48550/arXiv.2402.12880","url":null,"abstract":"There has been a range of studies of how autism is displayed in voice, speech, and language. We analyse studies from the biomedical, as well as the psychological domain, but also from the NLP domain in order to find linguistic, prosodic and acoustic cues. Our survey looks at all three domains. We define autism and which comorbidities might influence the correct detection of the disorder. We especially look at observations such as verbal and semantic fluency, prosodic features, but also disfluencies and speaking rate. We also show word-based approaches and describe machine learning and transformer-based approaches both on the audio data as well as the transcripts. Lastly, we conclude, while there already is a lot of research, female patients seem to be severely under-researched. Also, most NLP research focuses on traditional machine learning methods instead of transformers. Additionally, we were unable to find research combining both features from audio and transcripts.","PeriodicalId":508951,"journal":{"name":"Findings","volume":"55 6","pages":"1115-1125"},"PeriodicalIF":0.0,"publicationDate":"2024-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140446059","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
FindingsPub Date : 2024-02-19DOI: 10.48550/arXiv.2402.12267
Michela Lorandi, Anya Belz
{"title":"High-quality Data-to-Text Generation for Severely Under-Resourced Languages with Out-of-the-box Large Language Models","authors":"Michela Lorandi, Anya Belz","doi":"10.48550/arXiv.2402.12267","DOIUrl":"https://doi.org/10.48550/arXiv.2402.12267","url":null,"abstract":"The performance of NLP methods for severely under-resourced languages cannot currently hope to match the state of the art in NLP methods for well resourced languages. We explore the extent to which pretrained large language models (LLMs) can bridge this gap, via the example of data-to-text generation for Irish, Welsh, Breton and Maltese. We test LLMs on these under-resourced languages and English, in a range of scenarios. We find that LLMs easily set the state of the art for the under-resourced languages by substantial margins, as measured by both automatic and human evaluations. For all our languages, human evaluation shows on-a-par performance with humans for our best systems, but BLEU scores collapse compared to English, casting doubt on the metric’s suitability for evaluating non-task-specific systems. Overall, our results demonstrate the great potential of LLMs to bridge the performance gap for under-resourced languages.","PeriodicalId":508951,"journal":{"name":"Findings","volume":"86 3","pages":"1451-1461"},"PeriodicalIF":0.0,"publicationDate":"2024-02-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140449419","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
FindingsPub Date : 2024-02-13DOI: 10.48550/arXiv.2402.08155
Lin Tian, Xiuzhen Zhang, Jey Han Lau
{"title":"CMA-R: Causal Mediation Analysis for Explaining Rumour Detection","authors":"Lin Tian, Xiuzhen Zhang, Jey Han Lau","doi":"10.48550/arXiv.2402.08155","DOIUrl":"https://doi.org/10.48550/arXiv.2402.08155","url":null,"abstract":"We apply causal mediation analysis to explain the decision-making process of neural models for rumour detection on Twitter.Interventions at the input and network level reveal the causal impacts of tweets and words in the model output.We find that our approach CMA-R – Causal Mediation Analysis for Rumour detection – identifies salient tweets that explain model predictions and show strong agreement with human judgements for critical tweets determining the truthfulness of stories.CMA-R can further highlight causally impactful words in the salient tweets, providing another layer of interpretability and transparency into these blackbox rumour detection systems. Code is available at: https://github.com/ltian678/cma-r.","PeriodicalId":508951,"journal":{"name":"Findings","volume":"184 ","pages":"1667-1675"},"PeriodicalIF":0.0,"publicationDate":"2024-02-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140457348","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
FindingsPub Date : 2024-02-12DOI: 10.32866/001c.89707
Costanza Rossi, Fabian Reitemeyer, Oliver Heidrich, D. Rybski
{"title":"Comparison of Embodied Carbon of 3D-printed vs. Conventionally Built Houses","authors":"Costanza Rossi, Fabian Reitemeyer, Oliver Heidrich, D. Rybski","doi":"10.32866/001c.89707","DOIUrl":"https://doi.org/10.32866/001c.89707","url":null,"abstract":"The construction sector makes a considerable contribution to global greenhouse gas emissions. Considering the ongoing urbanization trends and climate change urgency, the exploration of alternative construction techniques should be a mandate. 3D-printing represents an emergent technology and more and more specimen are being built. We collect data of raw material use for houses that have already been built using 3D-printing. Assessing the construction related emissions, we find that the four examples, for which we could obtain the data, do have less emissions per square meter than conventionally built houses (10 international examples). We argue that 3D-printing represents an interesting alternative, but further research is necessary, not just in terms of environmental implications but also to better understand the social implications, e.g. health and safety or labor.","PeriodicalId":508951,"journal":{"name":"Findings","volume":"73 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-02-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139843987","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
FindingsPub Date : 2024-02-12DOI: 10.32866/001c.89707
Costanza Rossi, Fabian Reitemeyer, Oliver Heidrich, D. Rybski
{"title":"Comparison of Embodied Carbon of 3D-printed vs. Conventionally Built Houses","authors":"Costanza Rossi, Fabian Reitemeyer, Oliver Heidrich, D. Rybski","doi":"10.32866/001c.89707","DOIUrl":"https://doi.org/10.32866/001c.89707","url":null,"abstract":"The construction sector makes a considerable contribution to global greenhouse gas emissions. Considering the ongoing urbanization trends and climate change urgency, the exploration of alternative construction techniques should be a mandate. 3D-printing represents an emergent technology and more and more specimen are being built. We collect data of raw material use for houses that have already been built using 3D-printing. Assessing the construction related emissions, we find that the four examples, for which we could obtain the data, do have less emissions per square meter than conventionally built houses (10 international examples). We argue that 3D-printing represents an interesting alternative, but further research is necessary, not just in terms of environmental implications but also to better understand the social implications, e.g. health and safety or labor.","PeriodicalId":508951,"journal":{"name":"Findings","volume":"59 5","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-02-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139783758","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
FindingsPub Date : 2024-02-02DOI: 10.32866/001c.92735
Angela Curl, Anna Coppens, Cushla Dares, Jonathan Williman, H. Fitt, Simon Kingham
{"title":"Accessibility and Affordability Impacts of Half Price Public Transport Fares in Aotearoa New Zealand","authors":"Angela Curl, Anna Coppens, Cushla Dares, Jonathan Williman, H. Fitt, Simon Kingham","doi":"10.32866/001c.92735","DOIUrl":"https://doi.org/10.32866/001c.92735","url":null,"abstract":"Transport costs can constitute a large proportion of household budgets and restrict travel. Reduced or free public transport fares are one strategy to address these affordability and accessibility challenges. We analyse data from a 2022 transport survey to investigate participants perspectives and experiences of half price public transport fares in Ōtautahi Christchurch, a city in Aotearoa New Zealand. We find that those experiencing transport poverty are likely to have benefitted in terms of widening access and reduced financial stress.","PeriodicalId":508951,"journal":{"name":"Findings","volume":"39 6","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-02-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139809248","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
FindingsPub Date : 2024-02-02DOI: 10.32866/001c.92735
Angela Curl, Anna Coppens, Cushla Dares, Jonathan Williman, H. Fitt, Simon Kingham
{"title":"Accessibility and Affordability Impacts of Half Price Public Transport Fares in Aotearoa New Zealand","authors":"Angela Curl, Anna Coppens, Cushla Dares, Jonathan Williman, H. Fitt, Simon Kingham","doi":"10.32866/001c.92735","DOIUrl":"https://doi.org/10.32866/001c.92735","url":null,"abstract":"Transport costs can constitute a large proportion of household budgets and restrict travel. Reduced or free public transport fares are one strategy to address these affordability and accessibility challenges. We analyse data from a 2022 transport survey to investigate participants perspectives and experiences of half price public transport fares in Ōtautahi Christchurch, a city in Aotearoa New Zealand. We find that those experiencing transport poverty are likely to have benefitted in terms of widening access and reduced financial stress.","PeriodicalId":508951,"journal":{"name":"Findings","volume":"31 2","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-02-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139869036","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}