More Like this

1. Lifelong Learning Is an Ethical Responsibility of Professional Engineers: Is School Preparing Young Engineers for Lifelong Learning?

   https://doi.org/10.1061/(ASCE)EI.2643-9115.0000045  

   Froehle, Kamryn ; Phillips, Adam R. ; Murzi, Homero ( July 2021 , Journal of Civil Engineering Education)

   null (Ed.)
2. Lifelong DP: Consistently Bounded Differential Privacy in Lifelong Machine Learning

   Lai, Phung ; Hu, Han ; Phan, NhatHai ; Jin, Ruoming ; Thai, My T. ; Chen, An ( January 2022 , Conference on Lifelong Learning Agents. PMLR.)
3. Lifelong DP: Consistently Bounded Differential Privacy in Lifelong Machine Learning

   Lai, Phung ; Hu, Han ; Phan, Hai ; Jin, Ruoming ; Thai, My ; Chen, An ( August 2023 , Proceedings of The 1st Conference on Lifelong Learning Agents)
4. LIFELONG DP: CONSISTENTLY BOUNDED DIFFERENTIAL PRIVACY IN LIFELONG MACHINE LEARNING

   Lai, Phung ; Hu, Han ; Phan, NhatHai ; Jin, Ruoming ; Thai, My T. ; Chen, An M. ( July 2022 , 1st Conference on Lifelong Learning Agents)

   Precup, Doina ; Chandar, Sarath ; Pascanu, Razvan (Ed.)

   In this paper, we show that the process of continually learning new tasks and memorizing previous tasks introduces unknown privacy risks and challenges to bound the privacy loss. Based upon this, we introduce a formal definition of Lifelong DP, in which the participation of any data tuples in the training set of any tasks is protected, under a consistently bounded DP protection, given a growing stream of tasks. A consistently bounded DP means having only one fixed value of the DP privacy budget, regardless of the number of tasks. To preserve Lifelong DP, we propose a scalable and heterogeneous algorithm, called L2DP-ML with a streaming batch training, to efficiently train and continue releasing new versions of an L2M model, given the heterogeneity in terms of data sizes and the training order of tasks, without affecting DP protection of the private training set. An end-to-end theoretical analysis and thorough evaluations show that our mechanism is significantly better than baseline approaches in preserving Lifelong DP. The implementation of L2DP-ML is available at: https://github.com/haiphanNJIT/PrivateDeepLearning. 

   more » « less
5. Mutations in coral soma and sperm imply lifelong stem cell renewal and cell lineage selection

   https://doi.org/10.1098/rspb.2022.1766  

   López-Nandam, Elora H. ; Albright, Rebecca ; Hanson, Erik A. ; Sheets, Elizabeth A. ; Palumbi, Stephen R. ( January 2023 , Proceedings of the Royal Society B: Biological Sciences)

   In many animals, the germline differentiates early in embryogenesis, so only mutations that accumulate in germ cells are inherited by offspring. Exceptions to this developmental process may indicate other mechanisms have evolved to limit the effects of deleterious mutation accumulation. Stony corals are animals that can live for hundreds of years and have been thought to produce gametes from somatic tissue. To clarify conflicting evidence about germline-soma distinction in corals, we sequenced high coverage, full genomes with technical replicates for parent coral branches and their sperm pools. We identified post-embryonic single nucleotide variants (SNVs) unique to each parent branch, then checked if each SNV was shared by the respective sperm pool. Twenty-six per cent of post-embryonic SNVs were shared by the sperm and 74% were not. We also identified germline SNVs, those that were present in the sperm but not in the parent. These data suggest that self-renewing stem cells differentiate into germ and soma throughout the adult life of the colony, with SNV rates and patterns differing markedly in stem, soma and germ lineages. In addition to informing the evolution of germlines in metazoans, these insights inform how corals may generate adaptive diversity necessary in the face of global climate change.

    

   more » « less