https://scholar.dgist.ac.kr/handle/20.500.11750/56964 2026-04-04T12:19:06Z https://scholar.dgist.ac.kr/handle/20.500.11750/59222 Title: Stable olfactory receptor activation across odor complexity Author(s): Kim, Minseok; Lee, Jeongyoon; Park, Inah; Kim, Jihoon; Lee, Keunsoon; So, Jinhyun; Choi, Ji-Woong; Jang, Jae Eun; Kwon, Hyuk-Jun; Moon, Cheil; Choe, Han Kyoung Abstract: Mechanisms underlying single odorant activation of specific olfactory receptors are well understood. However, how the olfactory system processes complex odor mixtures at the receptor level remains unclear. This study examined olfactory receptor activation patterns across odor complexities using phosphoTRAP analysis. For most mixtures, receptor activation patterns closely matched the linear sum of individual component responses. However, distinct receptor sets display non-linear responses unexplained by linear models. Mixture responses were generally located between component responses and often aligned with linear predictions, though some deviations indicated non-linear interactions. Total activated receptors remained relatively constant regardless of odor complexity, suggesting efficient coding that prevented receptor saturation as odorant components increased. These findings provide receptor-level evidence that the olfactory system encodes complex odors primarily through linear integration of receptor activity, with added specificity from non-linear responses in limited receptors, advancing understanding of how the olfactory system normalizes receptor activation in response to natural odors. 2025-10-31T15:00:00Z https://scholar.dgist.ac.kr/handle/20.500.11750/56989 Title: LightVeriFL: A Lightweight and Verifiable Secure Aggregation for Federated Learning Author(s): Buyukates, Baturalp; So, Jinhyun; Mahdavifar, Hessam; Avestimehr, Salman Abstract: Secure aggregation protects the local models of the users in federated learning, by not allowing the server to obtain any information beyond the aggregate model at each iteration. Naively implementing secure aggregation fails to protect the integrity of the aggregate model in the possible presence of a malicious server forging the aggregation result, which motivates verifiable aggregation in federated learning. Existing verifiable aggregation schemes either have a linear complexity in model size or require time-consuming reconstruction at the server, that is quadratic in the number of users, in case of likely user dropouts. To overcome these limitations, we propose LightVeriFL, a lightweight and communication-efficient secure verifiable aggregation protocol, that provides the same guarantees for verifiability against a malicious server, data privacy, and dropout-resilience as the state-of-the-art protocols without incurring substantial communication and computation overheads. The proposed LightVeriFL protocol utilizes homomorphic hash and commitment functions of constant length, that are independent of the model size, to enable verification at the users. In case of dropouts, LightVeriFL uses a one-shot aggregate hash recovery of the dropped-out users, instead of a one-by-one recovery, making the verification process significantly faster than the existing approaches. Comprehensive experiments show the advantage of LightVeriFL in practical settings. IEEE 2024-03-31T15:00:00Z