It is part of the California Community Colleges System and the Peralta Community College District. Berkeley City College is accredited by the Accrediting Commission for Community and Junior Colleges.
Swarm robotics enables groups of simple robots to collaboratively perform complex tasks through local interactions and self-organisation. While centralised and rulebased swarm systems have shown promising results, they suffer from scalability limitations, single-point failures, and poor adaptability in dynamic environments. This paper presents a fully decentralised swarm robotics framework that integrates artificial intelligence-driven decision making using multi-agent reinforcement learning. Each robot operates as an autonomous agent relying solely on local observations and limited peer-to-peer communication, eliminating the need for global state information or centralised control. The proposed framework employs Independent Proximal Policy Optimization (IPPO) for decentralized policy learning under partial observability. Extensive simulation-based evaluation demonstrates that the proposed approach significantly outperforms rule-based and centralised baselines, achieving up to a $38 \%$ reduction in task completion time and a $25 \%$ improvement in fault tolerance under robot failure scenarios. Experiments were conducted using the PyBullet simulator with 1000 training episodes and local peer-to-peer communication constraints. The results confirm that decentralised learning-based swarm intelligence provides a scalable, resilient, and efficient solution for autonomous multi-robot coordination in complex environments.
The genus Myotis is one of the largest clades of bats, and it exhibits some of the most extreme variation in lifespans among mammals, alongside unique adaptations to viral tolerance and immune defence1-3. Here, to study the evolution of these phenotypes, we generated cell lines and near-complete genome assemblies for eight closely related Myotis species. Using genome-wide screens of positive selection, analyses of structural variation and functional experiments in primary cells, we identify patterns of adaptation contributing to longevity, cancer resistance and viral interactions. We demonstrate distinct modes of adaptation to DNA and RNA viruses compared with all other mammals, with bats exhibiting genome-wide over-representation of positive selection for DNA-virus-interacting proteins and elevated rates of copy-number variation for RNA-virus-interacting proteins. Characterization of Myotis-specific duplications of the key immune factor EIF2AK2 (also known as PKR) reveals multiple ancient segregating trans-species copy-number polymorphisms. We show that the recurrent evolution of longevity seen in Myotis is associated with positive selection in cancer pathways, and demonstrate a unique response to DNA damage in primary cells of the long-lived Myotis lucifugus. Together, our results suggest that bats' remarkable longevity and immunity are linked through pleiotropic adaptations to viruses and ageing-related disease.
Review of: Spaghetti Sissies: Queering Italian American Media , Julia Heim and Sole Anatrone (eds) (2023) Basingstoke: Palgrave Macmillan, 284 pp., ISBN 978-3-03110-196-0, h/bk, USD 139.99
Stem cell research offers unique opportunities for authentic scientific engagement, yet infrastructure requirements have confined participation to elite institutions, perpetuating workforce disparities. We developed an integrated framework combining stem cell engineering, cloud-connected microscopy, and psychometric assessment. The framework integrates a doxycycline-inducible NGN2 mouse embryonic stem cell (mESC) system, low-cost cloud microscopy, and the Stem Cell Research Identity Scale (SCRIS). Implementation across a high school and a community college demonstrated significant increases in scientific identity. Students using differentiating PSCs showed broader science identity gains than those using neuroblastoma cells, particularly in competence, research readiness, and recognition. High school students showed enhanced research competence gains compared to community college students despite equivalent intervention duration. Demographic analyses revealed enhanced effectiveness for Hispanic and first-generation college students. This framework provides a scalable model for broadening participation in biomedical research.
High-level behaviors such as sycophancy and hallucination correspond to directions in a model's activation space (Chen et al., 2025). These directions, known as persona vectors, can suppress undesirable traits either at inference time, subtracting the vector during generation, or through preventative steering, adding the vector during fine-tuning, so the optimizer never needs to encode the trait in weights, then removing the vector at evaluation. Chen et al. (2025) recommend preventative steering because it preserves Massive Multitask Language Understanding (MMLU) accuracy, but validations of this recommendation have measured capability alone. We primarily evaluate results on preventative steering of sycophancy and hallucination vectors on three models, measuring three safety benchmarks alongside trait expression, coherence, and MMLU. Our findings show that preventative steering still changes refusal behavior, but by an amount that depends on the trait and does not establish that preventative persona steering is safety-neutral, and steering methods should be evaluated on several safety benchmarks.