Pub Date : 2024-12-16DOI: 10.1038/s42256-024-00952-1
Nik Dennler, André van Schaik, Michael Schmuker
{"title":"Limitations in odour recognition and generalization in a neuromorphic olfactory circuit","authors":"Nik Dennler, André van Schaik, Michael Schmuker","doi":"10.1038/s42256-024-00952-1","DOIUrl":"10.1038/s42256-024-00952-1","url":null,"abstract":"","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"6 12","pages":"1451-1453"},"PeriodicalIF":18.8,"publicationDate":"2024-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142825245","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Survival analysis aims to estimate the impact of covariates on the expected time until an event occurs, which is broadly utilized in disciplines such as life sciences and healthcare, substantially influencing decision-making and improving survival outcomes. Existing methods, usually assuming similar training and testing distributions, nevertheless face challenges with real-world varying data sources, creating unpredictable shifts that undermine their reliability. This urgently necessitates that survival analysis methods should utilize stable features across diverse cohorts for predictions, rather than relying on spurious correlations. To this end, we propose a stable Cox model with theoretical guarantees to identify stable variables, which jointly optimizes an independence-driven sample reweighting module and a weighted Cox regression model. Through extensive evaluation on simulated and real-world omics and clinical data, stable Cox not only shows strong generalization ability across diverse independent test sets but also stratifies the subtype of patients significantly with the identified biomarker panels. Survival prediction models used in healthcare usually assume that training and test data share a similar distribution, which is not true in real-world settings. Cui and colleagues develop a stable Cox regression model that can identify stable variables for predicting survival outcomes under distribution shifts.
{"title":"Stable Cox regression for survival analysis under distribution shifts","authors":"Shaohua Fan, Renzhe Xu, Qian Dong, Yue He, Cheng Chang, Peng Cui","doi":"10.1038/s42256-024-00932-5","DOIUrl":"10.1038/s42256-024-00932-5","url":null,"abstract":"Survival analysis aims to estimate the impact of covariates on the expected time until an event occurs, which is broadly utilized in disciplines such as life sciences and healthcare, substantially influencing decision-making and improving survival outcomes. Existing methods, usually assuming similar training and testing distributions, nevertheless face challenges with real-world varying data sources, creating unpredictable shifts that undermine their reliability. This urgently necessitates that survival analysis methods should utilize stable features across diverse cohorts for predictions, rather than relying on spurious correlations. To this end, we propose a stable Cox model with theoretical guarantees to identify stable variables, which jointly optimizes an independence-driven sample reweighting module and a weighted Cox regression model. Through extensive evaluation on simulated and real-world omics and clinical data, stable Cox not only shows strong generalization ability across diverse independent test sets but also stratifies the subtype of patients significantly with the identified biomarker panels. Survival prediction models used in healthcare usually assume that training and test data share a similar distribution, which is not true in real-world settings. Cui and colleagues develop a stable Cox regression model that can identify stable variables for predicting survival outcomes under distribution shifts.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"6 12","pages":"1525-1541"},"PeriodicalIF":18.8,"publicationDate":"2024-12-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.nature.com/articles/s42256-024-00932-5.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142815895","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-13DOI: 10.1038/s42256-024-00943-2
Julian Büchel, Giacomo Camposampiero, Athanasios Vasilopoulos, Corey Lammie, Manuel Le Gallo, Abbas Rahimi, Abu Sebastian
Kernel functions are vital ingredients of several machine learning (ML) algorithms but often incur substantial memory and computational costs. We introduce an approach to kernel approximation in ML algorithms suitable for mixed-signal analogue in-memory computing (AIMC) architectures. Analogue in-memory kernel approximation addresses the performance bottlenecks of conventional kernel-based methods by executing most operations in approximate kernel methods directly in memory. The IBM HERMES project chip, a state-of-the-art phase-change memory-based AIMC chip, is utilized for the hardware demonstration of kernel approximation. Experimental results show that our method maintains high accuracy, with less than a 1% drop in kernel-based ridge classification benchmarks and within 1% accuracy on the long-range arena benchmark for kernelized attention in transformer neural networks. Compared to traditional digital accelerators, our approach is estimated to deliver superior energy efficiency and lower power consumption. These findings highlight the potential of heterogeneous AIMC architectures to enhance the efficiency and scalability of ML applications. A kernel approximation method that enables linear-complexity attention computation via analogue in-memory computing (AIMC) to deliver superior energy efficiency is demonstrated on a multicore AIMC chip.
{"title":"Kernel approximation using analogue in-memory computing","authors":"Julian Büchel, Giacomo Camposampiero, Athanasios Vasilopoulos, Corey Lammie, Manuel Le Gallo, Abbas Rahimi, Abu Sebastian","doi":"10.1038/s42256-024-00943-2","DOIUrl":"10.1038/s42256-024-00943-2","url":null,"abstract":"Kernel functions are vital ingredients of several machine learning (ML) algorithms but often incur substantial memory and computational costs. We introduce an approach to kernel approximation in ML algorithms suitable for mixed-signal analogue in-memory computing (AIMC) architectures. Analogue in-memory kernel approximation addresses the performance bottlenecks of conventional kernel-based methods by executing most operations in approximate kernel methods directly in memory. The IBM HERMES project chip, a state-of-the-art phase-change memory-based AIMC chip, is utilized for the hardware demonstration of kernel approximation. Experimental results show that our method maintains high accuracy, with less than a 1% drop in kernel-based ridge classification benchmarks and within 1% accuracy on the long-range arena benchmark for kernelized attention in transformer neural networks. Compared to traditional digital accelerators, our approach is estimated to deliver superior energy efficiency and lower power consumption. These findings highlight the potential of heterogeneous AIMC architectures to enhance the efficiency and scalability of ML applications. A kernel approximation method that enables linear-complexity attention computation via analogue in-memory computing (AIMC) to deliver superior energy efficiency is demonstrated on a multicore AIMC chip.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"6 12","pages":"1605-1615"},"PeriodicalIF":18.8,"publicationDate":"2024-12-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142815893","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-13DOI: 10.1038/s42256-024-00962-z
Johannes Brandstetter
Tackling partial differential equations with machine learning solvers is a promising direction, but recent analysis reveals challenges with making fair comparisons to previous methods. Stronger benchmark problems are needed for the field to advance.
{"title":"Envisioning better benchmarks for machine learning PDE solvers","authors":"Johannes Brandstetter","doi":"10.1038/s42256-024-00962-z","DOIUrl":"10.1038/s42256-024-00962-z","url":null,"abstract":"Tackling partial differential equations with machine learning solvers is a promising direction, but recent analysis reveals challenges with making fair comparisons to previous methods. Stronger benchmark problems are needed for the field to advance.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"7 1","pages":"2-3"},"PeriodicalIF":18.8,"publicationDate":"2024-12-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142815847","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-13DOI: 10.1038/s42256-024-00955-y
Nicolas Rouleau, Michael Levin
Sharp distinctions often drawn between machine and biological intelligences have not tracked advances in the fields of developmental biology and hybrid robotics. We call for conceptual clarity driven by the science of diverse intelligences in unconventional spaces and at unfamiliar scales and embodiments that blur conventional categories.
{"title":"Discussions of machine versus living intelligence need more clarity","authors":"Nicolas Rouleau, Michael Levin","doi":"10.1038/s42256-024-00955-y","DOIUrl":"10.1038/s42256-024-00955-y","url":null,"abstract":"Sharp distinctions often drawn between machine and biological intelligences have not tracked advances in the fields of developmental biology and hybrid robotics. We call for conceptual clarity driven by the science of diverse intelligences in unconventional spaces and at unfamiliar scales and embodiments that blur conventional categories.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"6 12","pages":"1424-1426"},"PeriodicalIF":18.8,"publicationDate":"2024-12-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142815848","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-12DOI: 10.1038/s42256-024-00948-x
Fan Yang, Fang Wang, Longkai Huang, Linjing Liu, Junzhou Huang, Jianhua Yao
{"title":"Reply to: Deeper evaluation of a single-cell foundation model","authors":"Fan Yang, Fang Wang, Longkai Huang, Linjing Liu, Junzhou Huang, Jianhua Yao","doi":"10.1038/s42256-024-00948-x","DOIUrl":"10.1038/s42256-024-00948-x","url":null,"abstract":"","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"6 12","pages":"1447-1450"},"PeriodicalIF":18.8,"publicationDate":"2024-12-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142809749","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-12DOI: 10.1038/s42256-024-00949-w
Rebecca Boiarsky, Nalini M. Singh, Alejandro Buendia, Ava P. Amini, Gad Getz, David Sontag
{"title":"Deeper evaluation of a single-cell foundation model","authors":"Rebecca Boiarsky, Nalini M. Singh, Alejandro Buendia, Ava P. Amini, Gad Getz, David Sontag","doi":"10.1038/s42256-024-00949-w","DOIUrl":"10.1038/s42256-024-00949-w","url":null,"abstract":"","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"6 12","pages":"1443-1446"},"PeriodicalIF":18.8,"publicationDate":"2024-12-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142809746","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-10DOI: 10.1038/s42256-024-00954-z
Christian Montag, Michèle Finck
The EU Artificial Intelligence Act bans certain “subliminal techniques beyond a person’s consciousness”, but uses undefined legal terms. Interdisciplinary efforts are needed to ensure effective implementation of the legal text.
{"title":"Successful implementation of the EU AI Act requires interdisciplinary efforts","authors":"Christian Montag, Michèle Finck","doi":"10.1038/s42256-024-00954-z","DOIUrl":"10.1038/s42256-024-00954-z","url":null,"abstract":"The EU Artificial Intelligence Act bans certain “subliminal techniques beyond a person’s consciousness”, but uses undefined legal terms. Interdisciplinary efforts are needed to ensure effective implementation of the legal text.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"6 12","pages":"1415-1417"},"PeriodicalIF":18.8,"publicationDate":"2024-12-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142796788","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-09DOI: 10.1038/s42256-024-00946-z
Haopeng Yu, Heng Yang, Wenqing Sun, Zongyun Yan, Xiaofei Yang, Huakun Zhang, Yiliang Ding, Ke Li
The complex ‘language’ of plant RNA encodes a vast array of biological regulatory elements that orchestrate crucial aspects of plant growth, development and adaptation to environmental stresses. Recent advancements in foundation models (FMs) have demonstrated their unprecedented potential to decipher complex ‘language’ in biology. In this study, we introduced PlantRNA-FM, a high-performance and interpretable RNA FM specifically designed for plants. PlantRNA-FM was pretrained on an extensive dataset, integrating RNA sequences and RNA structure information from 1,124 distinct plant species. PlantRNA-FM exhibits superior performance in plant-specific downstream tasks. PlantRNA-FM achieves an F1 score of 0.974 for genic region annotation, whereas the current best-performing model achieves 0.639. Our PlantRNA-FM is empowered by our interpretable framework that facilitates the identification of biologically functional RNA sequence and structure motifs, including both RNA secondary and tertiary structure motifs across transcriptomes. Through experimental validations, we revealed translation-associated RNA motifs in plants. Our PlantRNA-FM also highlighted the importance of the position information of these functional RNA motifs in genic regions. Taken together, our PlantRNA-FM facilitates the exploration of functional RNA motifs across the complexity of transcriptomes, empowering plant scientists with capabilities for programming RNA codes in plants. Approaches are needed to explore regulatory RNA motifs in plants. An interpretable RNA foundation model is developed, trained on thousands of plant transcriptomes, which achieves superior performance in plant RNA biology tasks and enables the discovery of functional RNA sequence and structure motifs across transcriptomes.
{"title":"An interpretable RNA foundation model for exploring functional RNA motifs in plants","authors":"Haopeng Yu, Heng Yang, Wenqing Sun, Zongyun Yan, Xiaofei Yang, Huakun Zhang, Yiliang Ding, Ke Li","doi":"10.1038/s42256-024-00946-z","DOIUrl":"10.1038/s42256-024-00946-z","url":null,"abstract":"The complex ‘language’ of plant RNA encodes a vast array of biological regulatory elements that orchestrate crucial aspects of plant growth, development and adaptation to environmental stresses. Recent advancements in foundation models (FMs) have demonstrated their unprecedented potential to decipher complex ‘language’ in biology. In this study, we introduced PlantRNA-FM, a high-performance and interpretable RNA FM specifically designed for plants. PlantRNA-FM was pretrained on an extensive dataset, integrating RNA sequences and RNA structure information from 1,124 distinct plant species. PlantRNA-FM exhibits superior performance in plant-specific downstream tasks. PlantRNA-FM achieves an F1 score of 0.974 for genic region annotation, whereas the current best-performing model achieves 0.639. Our PlantRNA-FM is empowered by our interpretable framework that facilitates the identification of biologically functional RNA sequence and structure motifs, including both RNA secondary and tertiary structure motifs across transcriptomes. Through experimental validations, we revealed translation-associated RNA motifs in plants. Our PlantRNA-FM also highlighted the importance of the position information of these functional RNA motifs in genic regions. Taken together, our PlantRNA-FM facilitates the exploration of functional RNA motifs across the complexity of transcriptomes, empowering plant scientists with capabilities for programming RNA codes in plants. Approaches are needed to explore regulatory RNA motifs in plants. An interpretable RNA foundation model is developed, trained on thousands of plant transcriptomes, which achieves superior performance in plant RNA biology tasks and enables the discovery of functional RNA sequence and structure motifs across transcriptomes.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"6 12","pages":"1616-1625"},"PeriodicalIF":18.8,"publicationDate":"2024-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.nature.com/articles/s42256-024-00946-z.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142793852","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-06DOI: 10.1038/s42256-024-00931-6
Yasha Ektefaie, Andrew Shen, Daria Bykova, Maximillian G. Marin, Marinka Zitnik, Maha Farhat
Deep learning has made rapid advances in modelling molecular sequencing data. Despite achieving high performance on benchmarks, it remains unclear to what extent deep learning models learn general principles and generalize to previously unseen sequences. Benchmarks traditionally interrogate model generalizability by generating metadata- or sequence similarity-based train and test splits of input data before assessing model performance. Here we show that this approach mischaracterizes model generalizability by failing to consider the full spectrum of cross-split overlap, that is, similarity between train and test splits. We introduce SPECTRA, the spectral framework for model evaluation. Given a model and a dataset, SPECTRA plots model performance as a function of decreasing cross-split overlap and reports the area under this curve as a measure of generalizability. We use SPECTRA with 18 sequencing datasets and phenotypes ranging from antibiotic resistance in tuberculosis to protein–ligand binding and evaluate the generalizability of 19 state-of-the-art deep learning models, including large language models, graph neural networks, diffusion models and convolutional neural networks. We show that sequence similarity- and metadata-based splits provide an incomplete assessment of model generalizability. Using SPECTRA, we find that as cross-split overlap decreases, deep learning models consistently show reduced performance, varying by task and model. Although no model consistently achieved the highest performance across all tasks, deep learning models can, in some cases, generalize to previously unseen sequences on specific tasks. SPECTRA advances our understanding of how foundation models generalize in biological applications. Ektefaie and colleagues introduce the spectral framework for models evaluation (SPECTRA) to measure the generalizability of machine learning models for molecular sequences.
{"title":"Evaluating generalizability of artificial intelligence models for molecular datasets","authors":"Yasha Ektefaie, Andrew Shen, Daria Bykova, Maximillian G. Marin, Marinka Zitnik, Maha Farhat","doi":"10.1038/s42256-024-00931-6","DOIUrl":"10.1038/s42256-024-00931-6","url":null,"abstract":"Deep learning has made rapid advances in modelling molecular sequencing data. Despite achieving high performance on benchmarks, it remains unclear to what extent deep learning models learn general principles and generalize to previously unseen sequences. Benchmarks traditionally interrogate model generalizability by generating metadata- or sequence similarity-based train and test splits of input data before assessing model performance. Here we show that this approach mischaracterizes model generalizability by failing to consider the full spectrum of cross-split overlap, that is, similarity between train and test splits. We introduce SPECTRA, the spectral framework for model evaluation. Given a model and a dataset, SPECTRA plots model performance as a function of decreasing cross-split overlap and reports the area under this curve as a measure of generalizability. We use SPECTRA with 18 sequencing datasets and phenotypes ranging from antibiotic resistance in tuberculosis to protein–ligand binding and evaluate the generalizability of 19 state-of-the-art deep learning models, including large language models, graph neural networks, diffusion models and convolutional neural networks. We show that sequence similarity- and metadata-based splits provide an incomplete assessment of model generalizability. Using SPECTRA, we find that as cross-split overlap decreases, deep learning models consistently show reduced performance, varying by task and model. Although no model consistently achieved the highest performance across all tasks, deep learning models can, in some cases, generalize to previously unseen sequences on specific tasks. SPECTRA advances our understanding of how foundation models generalize in biological applications. Ektefaie and colleagues introduce the spectral framework for models evaluation (SPECTRA) to measure the generalizability of machine learning models for molecular sequences.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"6 12","pages":"1512-1524"},"PeriodicalIF":18.8,"publicationDate":"2024-12-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142782801","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：