From Benchmark Accuracy to Industrial Evidence: A Registered Systematic Map and Nested Experiment-Level Audit of Validation Credibility in AI-Based Rotating-Machinery Fault Diagnosis
Artificial intelligence (AI) methods for rotating-machinery diagnosis based on vibration and acoustic signals are often evaluated on benchmark splits that do not establish physical independence or industrial generalization. This registered study combined a PRISMA-aligned systematic map with a nested experiment-level audit of validation credibility. Searches identified 11,916 records, retained 7520 unique records, and placed 5159 reports in the full-text population. The operational 500-record Batch 005 block produced a non-probability 351-report full-text queue. A post-review eligibility-consistency check yielded 97 included studies, 13 exclusions, 241 unresolved reports, and 157 validation experiments. Within this access-influenced subset, 130 experiments (82.8%) remained at A0–A1 physical independence, 27 (17.2%) reached A2–A4, and 41 (26.1%) evaluated B2–B4 domain shifts. Under a narrower leakage definition, positive evidence supported High risk in 19 experiments (12.1%), whereas 93 (59.2%) were Unclear because reporting was insufficient. Seventy-seven experiments reported classification-like numeric headline scores, but metric types were heterogeneous and were stratified rather than pooled. Metadata diagnostics confirmed source-structured selection and access-related differences between resolved and unresolved reports. The Industrial Evidence Matrix provides a structured vocabulary for validation evidence; all proportions describe the nested audit, not the complete literature.
It is found that Llama 3 delivers comparable quality to leading language models such as GPT-4 on a plethora of tasks, and performs competitively with the state-of-the-art on image, video, and speech recognition tasks.
This paper investigates developing real-world perception for active agents, proposes Gibson Environment for this purpose, and showcases a set of perceptual tasks learned therein.
F. Xia, Amir Zamir, Zhi-Yang He et al.· 2018 IEEE/CVF Conference on...· 1,069 citations· ⚡112
GAOKAO-Bench is introduced, an intuitive benchmark that employs questions from the Chinese GAOKAO examination as test samples, including both subjective and objective questions that contribute a robust evaluation benchmark for future large language models and offers valuable insights into the advantages and limitations...
Xiaotian Zhang, Chun-yan Li, Yi Zong et al.· arXiv.org· 216 citations· ⚡17
An algorithm is proposed which leverages disentangled semantic factors to generate adversarial perturbation by altering controlled semantic attributes to fool the learner towards various "adversarial" targets.
Haonan Qiu, Chaowei Xiao, Lei Yang et al.· arXiv.org· 208 citations· ⚡17
This work proposes a HALO that directly maximizes the utility of generations instead of maximizing the log-likelihood of preferences, and suggests that the best loss depends on the inductive biases most appropriate for a given setting, an oft-overlooked consideration.
This work investigates the possibilities of using LLMs in a resume screening setting via a document retrieval framework that simulates job candidate selection and finds that the MTEs are biased, significantly favoring White-associated names in 85% of cases and female-associated names in only 11.1% of cases.