Oct 2026· Zenodo (CERN European Organization for Nuclear Research)
Philosophy and Theoretical Science
Abstract
Theories of consciousness seek to explain what makes a system conscious, but they do not necessarily explain what makes a later conscious system the same subject as an earlier one. This article identifies this distinction as the Subjective Continuity Problem and examines its implications for consciousness science and artificial minds. I assess five influential approaches—global neuronal workspace theory, integrated information theory, higher-order theories, predictive processing, and recurrent processing theory—and find that none provides an explicit criterion for subjective continuity. Three conceptual challenges—interruption, reconstruction, and replication—together with a novel control case, the lockstep swap, demonstrate that consciousness criteria alone do not determine whether a particular subject persists when memory, personality, behaviour, and functional organisation are preserved. I then develop the AI Continuity Test, which applies six routine computational operations—pause and resume, copy, fork, merge, rollback, and migration—to examine whether existing theories can determine the number and persistence of subjects across changes in computational state and physical substrate. The analysis distinguishes consciousness detection from continuity determination and identifies unresolved questions concerning subject individuation, branching, restoration, and the aggregation of welfare. Even if Parfitian reductionism is correct that personal identity involves no further fact beyond physical and psychological relations, explicit criteria remain necessary for consistently counting and tracking subjects in scientific, medical, legal, ethical, and artificial-intelligence contexts. I do not propose a mechanism or claim to solve the problem. Instead, I specify the conditions an adequate continuity criterion must address, establishing subjective continuity as a distinct explanatory challenge for theories of consciousness.
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.
This work shows that orders of magnitude enhancement in performance could be obtained by a combination of hardware improvements and tight quantum-HPC integration and introduces high-performance architectures for quantum-probabilistic computing with custom-designed accelerators to tackle today's industry-scale classical...
Masoud Mohseni, Artur Scherer, K. Johnson et al.· arXiv.org· 121 citations· ⚡9
This paper presents a comprehensive overview of the Ultralytics YOLO family, emphasizing architectural evolution, benchmarking, deployment, and emerging directions from YOLOv5 through YOLO27, and examines detection, segmentation, depth, classification, pose, oriented detection, tracking, export, quantization, and deplo...
With $2.1 million funding from Google.org, the open-source Public Transit Intelligence Hub will unify public transit monitoring, operations, and passenger communication.