UCSC NLP at SemEval-2026 Task 10: Boundary-Aware Span Extraction and RoBERTa Classification for Conspiracy Detection
arXiv:2607.05689v1 Announce Type: new Abstract: We present our systems for SemEval-2026 Task 10 (PsyCoMark), addressing conspiracy marker extraction (Subtask 1) and document-level conspiracy detection (Subtask 2). For marker extraction, we formulate the task as multi-label span classification over enumerated candidate spans, using IoU >= 0.95 positive labeling, hard-negative sampling, and containment-based non-maximum suppression (NMS) with boundary-aware span representations. Document classification is modeled independently using a sequence classifier with label smoothing and a stratified tra
The proliferation of misinformation and coordinated disinformation campaigns across digital platforms necessitates advanced AI solutions for detection and mitigation.
Sophisticated detection of conspiracy theories can help maintain information integrity, mitigate societal polarization, and inform policy responses to online harms.
The development of robust, boundary-aware span extraction and classification models offers a more nuanced approach to identifying and understanding the linguistic structures of conspiracy theories.
- · Social Media Platforms
- · Fact-checking Organizations
- · Public Institutions
- · AI Ethics Researchers
- · Misinformation Propagators
- · State-sponsored Disinformation Campaigns
Improved automated detection of misinformation content on digital platforms.
Potential for increased censorship debates and challenges to free speech principles as detection capabilities advance.
The development of adversarial AI for misinformation, creating an arms race between detection and obfuscation technologies.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.CL