Accuracy
nullnullnullnullnullnullnullnull: A Benchmark for Parsing Ambiguous Questions into Database Queries
In each case, the ambiguity persists even when the database context is provided.
Appendix A Distribution Shift in Graph-Structured Data Distribution shift appears when the joint distribution differs between source domain and target domain
In all, structure shift is unique to graph data due to the non-IID nature caused by node interconnections. In this section, we provide additional details about the datasets used in our benchmark. The labels categorize airports by activity levels, measured in terms of flights or passenger numbers. Each dataset's nodes symbolize papers, while edges reflect their citation relationships. Specifically, "ACMv9" (A) includes papers from ACM spanning 2000 to 2010, "Citationv1" (C) consists of DBLP collected between 2004 and 2008.
Trap-MID: Trapdoor-based Defense against Model Inversion Attacks
While existing defenses often rely on regularization techniques to reduce information leakage, they remain vulnerable to recent attacks.
Hyper-opinion Evidential Deep Learning for Out-of-Distribution Detection
Evidential Deep Learning (EDL), grounded in Evidence Theory and Subjective Logic (SL), provides a robust framework to estimate uncertainty for out-of-distribution (OOD) detection alongside traditional classification probabilities.
Dissect Black Box: Interpreting for Rule-Based Explanations in Unsupervised Anomaly Detection Yu Zhang, Ruoyu Li, Nengwu Wu, Qing Li, Xinhan Lin, Y ang Hu, Tao Li, Yong Jiang
Delineation (GBD) algorithm is employed to define boundaries within each segmented distribution, effectively delineating normal from anomalous data points.
Privacy Backdoors: Enhancing Membership Inference through Poisoning Pre-trained Models
In this paper, we introduce a new type of model backdoor: the privacy backdoor attack .
Conjugated Semantic Pool Improves OOD Detection with Pre-trained Vision-Language Models
Observing that the original semantic pool is comprised of unmodified specific class names, we correspondingly construct a conjugated semantic pool (CSP) consisting of modified superclass names, each serving as a cluster center for samples sharing similar properties across different categories.