EditorialUsing AI and Social Media Multimodal Content for Disaster Response and Management: Opportunities, Challenges, and Future Directions
Introduction
During disasters and emergencies, getting a quick understanding of the situation as it unfolds is a challenging task for formal response organizations. Members of the public, humanitarian organizations, and other concerned authorities search for pertinent information either to prevent the crisis or to help victims as early as possible. However, traditional approaches to gain situational awareness (i.e., understanding a bigger picture of the situation) are time-consuming and labor-intensive (Hiltz and Plotnick, 2013). Widespread adoption of non-traditional information sources such as Social Media (SM) platforms has created numerous opportunities to gather relevant information in a timely manner to improve disaster response. Research studies show that the general public uses SM platforms during disasters and reports critical information such as early warnings, cautions, damages to infrastructure such as roads, bridges, and buildings (Castillo, 2016; Imran, Castillo, Diaz, and Vieweg, 2015; Vieweg, Castillo, and Imran, 2014). Moreover, SM is considered as a potential source to perform urgent needs assessment of the affected population after a major disaster (Vieweg et al., 2014). Research has also shown that many real-world events, including crises, are first reported on social media compared to traditional media outlets such as TV and radio (Kalyanam, Quezada, Poblete, and Lanckriet, 2016).1
Despite these benefits, processing SM data is not trivial. The various types of content available on SM platforms such as text messages, images, videos, are considered noisy and less formal compared to standard Web data, such as news articles or images published with those articles. Therefore, data analysis techniques trained and built for standard Web data often do not work well on SM data. For instance, text messages on Twitter or Facebook cannot exceed a fixed length and thus contain shortened words such as “2moro” instead of “tomorrow” in addition to other issues, including misspellings and use of slang. Studies revealed that techniques, which are specifically trained and built to process SM content, seem to work better than the ones trained for the processing of standard and well-structured data (Singh and Kumari, 2016). However, this highlights the need for in-domain labeled data, which is usually scarce.
Despite the lack of labeled data, the crisis informatics research community has continued to show advancements in different SM data processing tasks. Among others, these include data parsing, classification, extraction, summarizing, ranking, and recommendation. Several of these research lines have been struggling to perform until the recent breakthroughs in the field of Artificial Intelligence (AI), especially deep learning, have shown promising results. Deep learning techniques have recently been used by both Natural Language Processing (NLP) and Computer Vision (CV) communities to show substantial performance improvements over traditional machine learning approaches for several data processing tasks.
This special issue provided an opportunity to researchers working on crisis informatics topics using NLP, CV, ML, and AI methods to publish their latest and novel research works. We focused on the use of AI and SM multimodal content for disaster response topics. In total, we received 13 papers, which were rigorously peer-reviewed. Based on the expert reviews, four articles were finally accepted and summarized in Section 5. In the next section, we highlight the applications of SM textual content and techniques aiming at different processing tasks. Section 3 highlights applications of SM imagery content and how the CV community is addressing various challenges for crisis response, followed by a discussion on multimodal learning as a potential future research direction in Section 4. We conclude the editorial in Section 6.
Section snippets
SM text processing: opportunities, challenges, and future directions
Many SM platforms provide access to their data through APIs. Textual data available on SM is different in many aspects from other Web-based sources such as news articles. Typically, these messages use less formal language. They may contain words from multiple languages. They may also have various grammar and spelling mistakes. The length of the messages varies greatly, so does their content. They are mostly unstructured, fuzzy, and short in length (often due to a platform’s restricted character
SM image processing: opportunities, challenges, and future directions
During emergencies, people may find it more convenient to capture the moment via images and share them online to inform others about impending hazards, damage to critical infrastructure, etc. (Liu et al., 2008). To that end, images can provide more detailed information about the severity and extent of damage, better understanding of shelter needs and quality, more accurate assessment of ongoing rescue operations, and faster identification of lost or injured people, among others. A survey study
Social media multimodal analysis
SM data have a multimodal nature, i.e., text messages, images, videos, and other meta-data appear together in general. These modalities oftentimes contain complementary information that, when analyzed jointly, can prove extremely useful to ascertain the big picture of a disaster situation at a greater level of detail (Alam, Ofli, and Imran, 2019; Alam, Ofli, Imran, and Aupetit, 2018c). For instance, a number of studies have shown the value of multimodal analysis as a better tool for relevancy
Special issue and accepted article summaries
In total, this special issue received 13 submissions. Of all, two were desk rejected. The remaining articles went through a rigorous reviewing process where all the articles were reviewed by at least three expert reviewers followed by a meta-review performed by one of the guest editors. Finally, four articles are accepted, which we summarize next.
Conclusion
SM during time-critical situations such as disasters and emergencies is considered as a vital resource for responders and decision-makers. Text messages and images shared on SM during such events contain situational as well as actionable information. Despite their usefulness, majority of this pertinent data is not available to humanitarian organizations during disasters, mainly due to several data processing and data quality challenges. This editorial provides a brief review of challenges and
References (114)
- et al.
A survey on automatic image caption generation
Neurocomputing
(2018) - et al.
A survey of approaches and trends in person re-identification
Image and vision computing
(2014) - et al.
Public behavior response analysis in disaster events utilizing visual analytics of microblog data
Computers & Graphics
(2014) - et al.
Harvesting social media for generation of near real-time flood maps
Procedia Engineering
(2016) - et al.
Identifying crisis-related informative tweets using learning on distributions
Information Processing & Management
(2020) - et al.
Rapid relevance classification of social media posts in disasters and emergencies: A system and evaluation featuring active, incremental and online learning
Information Processing & Management
(2020) - et al.
Crisis communication, learning and responding: Best practices in social media
Computers in Human Behavior
(2016) - et al.
Social media in emergency management: Twitter as a tool for communicating risks to the public
Technological Forecasting and Social Change
(2016) - et al.
Understanding communication dynamics on twitter during natural disasters: A case study of hurricane sandy
International journal of disaster risk reduction
(2019) - et al.
Emergency service’ attitudes towards social media: A quantitative and qualitative survey across europe
International Journal of Human-Computer Studies
(2016)
Role of text pre-processing in twitter sentiment analysis
Procedia Computer Science
Wildfire detection using transfer learning on augmented datasets
Expert Systems with Applications
Mining location from social media: A systematic review
Computers, Environment and Urban Systems
Crisis communication online: How medium, crisis type and emotions affected public reactions in the fukushima daiichi nuclear disaster
Public relations review
Crisis-dias: Towards multimodal damage analysis – deployment, challenges and assessment
Aaai conference on artificial intelligence: Ai for social impact track
A comparative study of global and deep features for the analysis of user-generated natural disaster related images
13th image, video, and multidimensional signal processing workshop (ivmsp)
Convolutional neural networks for disaster images retrieval
Proc. of the mediaeval workshop
Image4Act: Online social media image processing for disaster response
Ieee/acm international conference on advances in social networks analysis and mining (ASONAM)
CrisisMMD: Multimodal twitter datasets from natural disasters
International aaai conference on web and social media (icwsm)
Processing social media images by combining human and machine computing during crises
International Journal of Human–Computer Interaction
Descriptive and visual summaries of disaster events using artificial intelligence techniques: case studies of hurricanes harvey, irma, and maria
Behaviour & Information Technology
A twitter tale of three hurricanes: Harvey, irma, and maria
International conference on information systems for crisis response and management (iscram)
Can we predict a riot? disruptive event detection using twitter
ACM Transactions on Internet Technology (TOIT)
Unsupervised detection of sub-events in large scale disasters
arXiv preprint arXiv:1912.13332
Tweeting situational awareness during the sydney siege
Journal of Policing, Intelligence and Counter Terrorism
A deep learning approach to universal image manipulation detection using a new convolutional layer
Proceedings of the 4th acm workshop on information hiding and multimedia security
Image manipulation detection
Journal of Electronic Imaging
An overview of sentiment analysis in social media and its applications in disaster relief
Sentiment analysis and ontology engineering
Visual representations of disaster
Cscw
Global-local feature fusion for image classification of flood affected roads from social multimedia
Proc. of the mediaeval workshop
On semantics and deep learning for event detection in crisis situations
Eswc workshop on semantic deep learning (semdeep)
Illuminant-based transformed spaces for image forensics
IEEE Transactions on Information Forensics and Security
Big crisis data: social media in disasters and time-critical situations
Information credibility on twitter
Proceedings of the 20th international conference on world wide web
Diverse and coherent paragraph generation from images
The european conference on computer vision (eccv)
Understanding and classifying image tweets
Acm international conference on multimedia
The placing task: A large-scale geo-estimation challenge for social-media videos and images
Proceedings of the 3rd acm multimedia workshop on geotagging and its applications in multimedia
Mining and classifying image posts on social media to analyse fires
International conference on information systems for crisis response and management (iscram)
Supporting disaster reconnaissance with social media data: A design-oriented case study of the 2013 colorado floods
International conference on information systems for crisis response and management (iscram)
Towards understanding crisis events on online social networks through pictures
Ieee/acm international conference on advances in social networks analysis and mining (asonam)
What makes paris look like paris?
ACM Transactions on Graphics (SIGGRAPH)
The deepfake detection challenge (dfdc) preview dataset
arXiv preprint arXiv:1910.08854
Social sensing in disaster city digital twin: Integrated textual–visual–geo framework for situational awareness during built environment disruptions
Journal of Management in Engineering
Vost: A case study in voluntary digital participation for collaborative emergency management
Information Processing & Management
Ensembled convolutional neural network models for retrieving flood relevant tweets
Proc. of the mediaeval workshop
Multimodal analysis of disaster tweets
Fifth international conference on multimedia big data (bigmm)
Faking sandy: characterizing and identifying fake images on twitter during hurricane sandy
Proceedings of the 22nd international conference on world wide web
Im2gps: estimating geographic information from a single image
The ieee conference on computer vision and pattern recognition (cvpr)
The signals and noise: actionable information in improvised social media channels during a disaster
Proceedings of the 2017 acm on web science conference
Dealing with information overload when using social media for emergency management: Emerging solutions
International conference on information systems for crisis response and management (iscram)
Cited by (98)
Unveiling the dynamics of crisis events: Sentiment and emotion analysis via multi-task learning with attention mechanism and subject-based intent prediction
2024, Information Processing and ManagementInvestigating the emotional experiences in eSports spectatorship: The case of League of Legends
2023, Information Processing and ManagementCross-modal fine-grained alignment and fusion network for multimodal aspect-based sentiment analysis
2023, Information Processing and ManagementLocating and deploying essential goods and equipment in disasters using AI-enabled approaches: A systematic literature review
2023, Progress in Disaster ScienceSSM: Stylometric and semantic similarity oriented multimodal fake news detection
2023, Journal of King Saud University - Computer and Information SciencesPublic opinion changing patterns under the double-hazard scenario of natural disaster and public health event
2023, Information Processing and Management