Paper deep dive
Life Operators: a self-evolving framework for multiscale life modelling
Shuo Wang, Yike Guo
Intelligence
Status: succeeded | Model: Gemma-4-26B-A4B | Prompt: intel-v1 | Confidence: 93%
Last extracted: 9/2/2026, 3:59:18 AM
Summary
The paper introduces 'Life Operators', a self-evolving framework for multiscale life modelling that unifies statistical and mechanistic approaches. It defines three core operator roles—Perception (inferring states from observations), Evolution (propagating states under dynamics/interventions), and Generation (mapping states to signals)—connected by Bridge operators. These form task-specific Operator Graphs. The framework supports evidence-gated self-evolution, where an AI co-scientist proposes revisions to the model components, which are then validated against independent evidence to update the model library.
Entities (13)
Relation Signals (12)
Yike Guo → affiliatedwith → Hong Kong University of Science and Technology
confidence 99% · Yike Guo 2 ... 2 Hong Kong University of Science and Technology
Shuo Wang → affiliatedwith → Fudan University
confidence 99% · Shuo Wang 1 ... 1 Digital Medical Research Center, Fudan University
Life Operators → includes → Evolution Operator
confidence 95% · Evolution operators propagate these states under natural or intervention-conditioned dynamics
Life Operators → includes → Generation Operator
confidence 95% · Generation operators map them to measurable signals.
Life Operators → includes → Perception Operator
confidence 95% · We propose Life Operators: task-bounded mappings that define three scientific roles. Perception operators...
Life Operators → includes → Bridge Operator
confidence 92% · Bridge operators connect components with different variables, scales and time steps.
Perception Operator → infers → biological states
confidence 90% · Perception operators infer task-relevant biological states from multimodal observations
Generation Operator → maps → biological states
confidence 90% · Generation operators map them to measurable signals.
Cypher Suggestions (0)
No Cypher suggestions yet.
Abstract
Abstract:Medical AI is moving beyond recognition towards clinical dialogue and longitudinal prediction. Yet a central question remains: how would a patient's state change under intervention? Statistical models learn future observations, whereas mechanistic models describe selected processes. Neither provides a common framework for representing patient state, coupling scales or revising failed assumptions. We propose Life Operators: task-bounded mappings that define three scientific roles. Perception operators infer task-relevant biological states from multimodal observations, Evolution operators propagate these states under natural or intervention-conditioned dynamics, and Generation operators map them to measurable signals. Each role may be realised by equations, statistical models, neural networks or hybrids. Bridge operators connect components with different variables, scales and time steps. Selected operators and bridges form task-specific Operator Graphs containing the smallest set of states and mechanisms sufficient for a declared claim. This modular structure also makes scientific revision localisable. An AI co-scientist may propose changes to states, operators, bridges or graph structure, while independent evidence determines which variants are retained, restricted or retired. Over time, validated components could accumulate into broader multiscale models of the human body and provide a computational foundation for medical artificial superintelligence.
Tags
Links
- Source: https://arxiv.org/abs/2609.00068v1
- Canonical: https://arxiv.org/abs/2609.00068v1
Trouble viewing inline? Open PDF directly →
Full Text
37,717 characters extracted from source content.
Expand or collapse full text
Life Operators: a self-evolving framework for multiscale life modelling Shuo Wang 1 Yike Guo 2 1 Digital Medical Research Center, Fudan University, Shanghai, China 2 Hong Kong University of Science and Technology, Hong Kong, China shuowang@fudan.edu.cn yikeguo@ust.hk Abstract Medical AI is moving beyond recognition towards clinical dialogue and longitudinal prediction. Yet a central question remains: how would a patient’s state change un- der intervention? Statistical models learn future observations, whereas mechanistic models describe selected processes. Neither provides a common framework for repre- senting patient state, coupling scales or revising failed assumptions. We proposeLife Operators: task-bounded mappings that define three scientific roles. Perception op- erators infer task-relevant biological states from multimodal observations, Evolution operators propagate these states under natural or intervention-conditioned dynamics, and Generation operators map them to measurable signals. Each role may be realised by equations, statistical models, neural networks or hybrids. Bridge operators connect components with different variables, scales and time steps. Selected operators and bridges form task-specific Operator Graphs containing the smallest set of states and mechanisms sufficient for a declared claim. This modular structure also makes scien- tific revision localisable. An AI co-scientist may propose changes to states, operators, bridges or graph structure, while independent evidence determines which variants are retained, restricted or retired. Over time, validated components could accumulate into broader multiscale models of the human body and provide a computational foundation for medical artificial superintelligence. 1 Introduction Medicine asks three progressively more demanding questions: What is happening now? What is likely to happen next under prevailing conditions? What would happen under a specified intervention? The first asks which biological states are consistent with the available observations. The second asks how those states are expected to change over time. The third compares action-dependent trajectories. For the same patient, watchful waiting, a drug, 1 arXiv:2609.00068v1 [cs.CL] 30 Aug 2026 infer the current statepredict the natural course What is happeningnow? What is likely tohappennext? compare intervention trajectories What wouldchange underintervention? Molecular DNA, RNA, proteins and metabolites CellularTissue cell–cell interactions and microenvironment Organ structure and function Individual whole-body state and phenotype Molecular therapiesSurgeryanddevice- basedtherapy Cell and gene therapiesTissue-directedtherapyBehavioural and digital interventions Medical questions Biological scales Representative interventions intracellular states and processes Figure 1:Medical questions, biological scales and representative interventions. Three medical questions are aligned with biological organization from molecular to individual scales and representative interventions acting at these levels. Bidirectional arrows indicate interactions across adjacent scales. ablation or a device may lead to different future states, clinical endpoints and paths between them. A recognition model may address the first question without representing temporal change. A forecasting model may address the second without specifying how an intervention enters the underlying biological process. Addressing the third requires a model in which interventions enter the biological dynamics explicitly. Figure 1relates these questions to biological scales and to representative interventions at different levels of the system. Data-driven medical AI is beginning to address a broader range of clinical problems. AI- assisted electrocardiography and echocardiography have entered prospective evaluation for detection and measurement within care pathways[1,2]. Conversational systems have pro- gressed from diagnostic dialogue to reasoning over multi-visit disease management[ 3,4]. Longitudinal models learn temporal structure from electronic health records, imaging, age and genetic risk[ 5–7]. Together, these developments extend medical AI from present-state recognition towards disease-trajectory modelling. Predicting a clinical record is not the same as modelling the biological process that generated it. Recorded events reflect both changes in the patient and the process of care. The latter includes testing, treatment assignment, follow-up and coding[8]. Predicting future records therefore does not, by itself, identify a task-relevant biological state or its transition law. Likewise, conditioning a forecast on an observed treatment label does not establish what would have occurred for the same patient under an alternative treatment. Such models may remain clinically useful without supporting mechanistic interpretation or individual counterfactual claims. Mechanistic models make a complementary trade-off: they provide explicit variables and 2 structural assumptions, but usually for selected processes and scales. Cellular electrophys- iology, reaction–diffusion models and systems biology have each made selected processes of biological change amenable to quantitative analysis[9–11]. Cardiac digital twins further illustrate how anatomical and physiological parameters can be inferred from patient obser- vations and used to parameterise simulations[12,13]. Yet fidelity within a local mechanism does not ensure patient specificity or valid composition across scales. Parameters may be weakly identifiable, and boundary conditions may be incompletely observed. Effective re- lations may also become unreliable when transferred to another scale or intervention. The central problem is therefore not whether data-driven or mechanistic models should prevail. The challenge is instead to make partial models answer one medical question together, even when they use different mathematical forms and sources of evidence. The assumptions of each model must remain explicit. We proposeLife Operatorsas a common language, not another model class. This language specifies what an executable component claims to infer, evolve or generate and states the assumptions and domain of that claim. Perception, Evolution and Generation denote these three scientific roles, while equations, statistical models, neural networks and hybrids are possible implementations. Bridge operators provide explicit mechanisms for composition, and task-specific graphs organise the components required by a medical question. This modular structure provides a basis for evidence-gated revision. Common roles do not imply common algorithms or equal evidential status. The immediate objective is not a complete digital replica of a person. We instead begin with a declared endpoint, time horizon, set of interventions and available observations. We then construct the smallest state and graph sufficient for that question. Whether independently validated operators and bridges can accumulate into broader models of the human body is an empirical question that must be tested through bounded medical problems. Medical artificial superintelligence is therefore a possible long-term consequence of such accumulation. 2 Life Operators Life as a dynamical system.For a specified medical problem, health and disease may be represented as an open, stochastic, controlled and partially observed dynamical system. A minimal formulation is s t+∆t =F ∆t (s t , u t , c t , ξ t ),x (m) t p ( x (m) t js t , a (m) t ) .(1) Here,tindexes time and∆tis the propagation interval. The termF ∆t is the corresponding state-transition map. The variables t denotes the task-relevant biological state,u t the inter- vention andc t the individual and environmental context. The termξ t captures unresolved influences and stochasticity. The observationx (m) t , wheremindexes modality, may be an im- age, a physiological signal, a genomic or transcriptomic measurement, single-cell sequencing data, a molecular assay or a clinical record. The acquisition variablea (m) t describes how the corresponding modality is acquired and recorded. It includes the device, protocol, sampling, detection and reporting processes. When the modality index is omitted below,x t anda t 3 bLifeasadynamicalsystem 푠 !"#! =퐹 #! 푠 ! ,푢 ! ,푐 ! ,휉 ! 푠 !"#! 푎 !"#! $ 푠 ! 푠 !%#! 푠 !%&#! 푎 ! $ 푎 !%#! $ 푎 !%&#! $ 푥 ! $ ∼푝푥 ! $ 푠 ! ,푎 ! $ Latentstate Observations Acquisition/ measurement 푢 ! 푐 ! interventioncontext 휉 ! unresolved Influences/noise omicsimagingphysiology biomarkers years days- months hours sec-min ns-ms MoleculeCellTissueOrganHuman Molecular Dynamics Virtual Cell Virtual tissue Organ Digital Twin Patient Digital Twin Life Operators across scales aUnifiedmultiscalelifemodelling cP-E-Glifeoperators Evolution(E)Generation(G)Perception(P) observation→ state state+intervention/context → futurestate state→ observation Bridge operator (C↔T) P E G Bridge operator (T↔O) P E G Bridge operator (O↔P) P E G celltissueorganpatient P E G Multiscaleoperatorlibrary dTask-specificoperatorgraphs P (cell) E (cell) Bridge (C↔T) E (tissue) Bridge (T↔O) E (organ) G (organ) P (tissue) P (organ) input data output data Activeoperator Inactiveoperator intervention Figure 2:Life Operators for multiscale modelling. a, Models spanning biological scales and characteristic timescales.b, Life as a partially observed dynamical system shaped by interventions, context and unresolved influences.c, A multiscale library of Perception, Evo- lution and Generation operators linked by bridge operators.d, A task-specific Operator Graph composing selected operators and bridges from input data and intervention to pre- dicted output; dashed nodes are inactive. denote the corresponding multimodal collections. The state is not intended to reproduce the complete microphysical configuration of a person or to compress all available data without loss. It is an abstraction chosen for a question. Lety t:t+τ denote the outcome trajectory of interest over horizonτ. Within a stated domain, an inferred state is sufficient if the observed past adds no material information about the outcome onces t , the allowed inputs and the context are known: p(y t:t+τ js t , x 1:t , u t:t+τ , c t:t+τ )'p(y t:t+τ js t , u t:t+τ , c t:t+τ ).(2) Figure2summarises a progression from multiscale biological models, through dynamical systems and P–E–G operator libraries, to task-specific Operator Graphs. Three operator roles.The partially observed system exposes three basic computational problems. The latent state must be inferred from observations, propagated under natural or intervention-conditioned dynamics, and related back to measurable signals. We define the 4 corresponding operator classes in parallel: P:(x 1:t , a 1:t , u 1:t−1 , c 1:t−1 )7!p(s t jx 1:t , a 1:t , u 1:t−1 , c 1:t−1 ), E ∆t :(s t , u t , c t )7!p(s t+∆t js t , u t , c t ), G:(s t , a t )7!p(x t js t , a t ).(3) Life Operators are defined by the scientific mappings they claim to perform, not by the algorithms used to implement them. Their implementations may include ordinary or partial differential equations, numerical solvers, statistical transition models and neural networks. Neural operators, learned latent dynamics and hybrid methods may also realise the same role. Mechanistic implementations encode selected structure. Learned components may approximate unresolved relations or accelerate repeated simulation and inversion[14–16]. ThePerception Operatormaps observations to a posterior distribution over task-relevant biological states. It captures what can be inferred from incomplete observations and the uncertainty of that inference. It therefore avoids treating a single reconstruction as ground truth. This posterior provides the state representation used by the other operators. TheEvolution Operatorspecifies a transition distribution from a current state to a future state over a defined time interval. This distribution is conditional on context and, where relevant, an explicitly defined intervention. It may represent continuous dynamics, stochastic transitions, discrete events or mixtures of these. Within the same dynamical framework, it can describe both natural progression and intervention-conditioned trajectories. TheGeneration Operatormaps a biological state to a distribution of measurable signals. It represents biological signal formation and, where relevant, the processes by which signals are acquired and recorded. In this way, it connects the internal state represented by the model to empirical observations. Inference and uncertainty.Bayesian inference connects the three roles without forming a fourth operator. For clarity, we use a simplified discrete-time filtering recursion that omits parameters and acquisition variables. In this recursion, Evolution supplies the transition prior, Generation the observation likelihood and Perception the posterior update: p(s t jx 1:t , u 1:t−1 , c 1:t−1 )/p(x t js t ) ∫ p(s t js t−1 , u t−1 , c t−1 )p(s t−1 jx 1:t−1 , u 1:t−2 , c 1:t−2 )ds t−1 . (4) This recursion separates change in the patient from change in knowledge about the patient. Natural variability, measurement error, parameter uncertainty and uncertainty about model form need not be collapsed into a single confidence score. 3 Operator Graphs Graphs defined by the question.Biological processes may be represented by compo- nents that differ in variables, scales, units and time steps. A task-specificLife Operator Graph(hereafter, Operator Graph) selects and connects the Perception, Evolution and Gen- 5 eration operators required for a medical question. The endpoint, time horizon, interventions and available observations determine which states, operators, bridges and scales the graph requires. For a declared questionQ, letΓ Q =(V Q , E Q ). Its nodesV Q are selected Perception, Evolution and Generation operators. Its edgesE Q are compatible interfaces or bridge oper- ators. The graph is therefore a testable hypothesis about which components are sufficient for the declared claim. Bridges between operators.The outputs of selected components are not automatically compatible, so each connection requires an explicit transformation. Abridge operatortrans- forms the output of one component into the typed input required by another. For a source componentiand a receiving componentj, the minimal relation is r i→j,t =B i→j ( z (i) t ) ,(5) wherez (i) t is the relevant source state or output andr i→j,t is the typed message passed to the receiver. The message may carry an effective parameter, source term, boundary condition or distribution together with its uncertainty. AScale Bridgeconnects components at different biological scales. Its upward and downward mappings have different meanings and are not assumed to be inverses. A bridge may be mechanistic, statistical or learned. Regardless of implementation, it must preserve the meaning, units, timing and uncertainty required by the receiving component. Computational compatibility or speed alone does not establish that the resulting Operator Graph is scientifically valid. Each operator and bridge therefore requires a concise scientific contract. This contract states the component’s role, inputs and outputs, variables, units and assumptions. It also records the valid domain, uncertainty, supporting evidence, version and failure conditions. Such contracts make components traceable and comparable, but interface compatibility remains only a prerequisite for composition. Validation before expansion.Validation must be specific to the receiving component and the declared task. A bridge should preserve the distinctions, uncertainty and intervention responses needed by the receiver. The resulting Operator Graph should be tested against independent observations. If validation fails, the graph may revise the message, narrow its domain or horizon, or activate a finer model. The graph should expand only when the evidence requires it. Missing observations call for additional measurement or wider uncertainty. A missing scale may require a new bridge or subgraph. Missing evidence for an intervention requires interventional evidence, not additional model complexity. Components that do not improve the declared prediction or its calibration should be removed. Box 1shows how a bounded question in hypertrophic cardiomyopathy (HCM) can be com- piled into explicit states, operators, bridges, outputs and falsification tests. 6 BOX 1 An example of building an Operator Graph for drug response in HCM This example compiles one prospective test rather than a complete HCM digital twin. For an adult with symptomatic obstructive HCM, it predicts the joint distribution of changes in Valsalva LVOT gradient (VLVOTg) and LVEF at a prespecified follow-up under a declared mavacamten regimen. The cited studies support cell-to-organ simulation, population exposure–response modelling and trial-average effects separately; they do not validate their integration into an individual-response graph[17–19]. Minimal candidate graph. x 0 P base !p(s 0 jx 0 ), (u 0:T ,c PK ) E PK !p(C 0:T ) E cell !p(T a ,τ r ), p(T a ,τ r ) B cell→organ !p(σ a ), (s 0 ,σ a ,c Val ) E organ !p(V ED ,V ES ,v LVOT ) G echo !p(∆VLVOTg,∆LVEF). Step 1: Infer the baseline state.Inputs x 0 include protocolised echocardiography, ventric- ular and mitral geometry, blood pressure, heart rate, rhythm and treatment history. Perception re- turns a posterior overs 0 = (g,ℓ 0 ,κ,α 0 ): geometry, loading, passive stiffness and baseline contractility. Dose and adherence defineu 0:T ; drug-metabolism factors and interacting treatments definec PK ; the Valsalva protocol and its loading context define c Val . Step 2: Propagate the intervention.Phar- macokinetic Evolution maps the regimen to con- centrationC 0:T ; cellular Evolution maps concen- tration to active tensionT a and relaxationτ r . The Scale Bridge maps these cellular outputs to fibre- oriented active stressσ a . Organ Evolution is the key candidate: it combines ventricular geometry, mitral-valve systolic anterior motion, loading and stress to generate pressure–volume dynamics and LVOT flow. It is proposed here, not treated as val- idated. Step 3: Recover a known response.Genera- tion applies the echo acquisition protocola echo and converts organ states to observables usingLVEF = 1V ES /V ED and the simplified Bernoulli relation VLVOTg4(v LVOT cosθ) 2 , whereθis the beam– flow angle, VLVOTg is in mmHg andv LVOT is in m s −1 . Before individual prediction, the graph should recover the observed population exposure– response relations for VLVOTg and LVEF. This positive control tests intervention response; base- line agreement does not. Step 4: Test each contract.Test pharmacoki- netics on held-out concentration profiles; cellular dynamics on concentration-matched human force and relaxation data; bridges and organ dynamics on independent deformation, pressure–volume and LVOT-flow observations under unseen loads; and Generation on blinded, protocol-matched echocar- diography. Step 5: Freeze, compare and ablate.Freeze the graph before external serial evaluation. Com- pare calibration, prespecified interval coverage and joint predictive score with a simpler exposure– response model, then ablate the cellular and or- gan components in turn. Retain a scale only if it improves external prediction, calibration or failure localisation. 4 Evidence-Gated Self-Evolution Modular and versioned revision.A modular specification represents a bounded biolog- ical process and supports systematic revision of the corresponding model. At revisionk, a versioned operator library is compiled into a task-specific graph. Each component declares its scientific role, interface and valid domain. A failure can therefore be traced to Perception, 7 ℳ ! =풮 ! ,Ω ! ,ℬ ! ,Γ ",! 1.Currentversionedlibraryandmodelℳ ! state definitions operator library bridge library graph structure PEG Questions: 푄 ! 푄 " 푄 # 2.Task-specificuse andmonitoring 3.New data and discrepancies 풟 $! %&' ∆ ! 4.Proposalgeneration(AIco-scientist) ∆풮 revisedstate ∆Ω revisedoperator ∆ℬ revisedbridge ∆Γ revisedgraph PEG hold-out data 5.Independentevidencegate perturbation experiments prospective validation real-world evidence ℳ ! →ℳ !() 6.Updated library and modelℳ !"# state definitions operator library bridge library graph structure PEG Questions: 푄 ! 푄 " 푄 # 퐶 !() =Propose *+ 풮 ! ,ℳ ! ,풟 $! %&' ,∆ ! AIacceleratesproposal:evidence-gated Figure 3:Evidence-gated self-evolution of a versioned Life Operator model.Task- specific use of the current modelM k produces new development data and exposes discrepan- cies. An AI co-scientist proposes revisions to state definitions, operator and bridge libraries, or graph structure. Held-out data, perturbation experiments, prospective validation and real-world evidence gate promotion to the updated modelM k+1 . AI accelerates proposal generation; independent evidence determines which changes persist. Evolution, Generation, a bridge or graph composition. The model need not be retrained as a single undifferentiated system. Components can then be added, replaced, restricted or retired while the rest of the graph remains fixed. Large historical and multimodal datasets can support the learning of reusable operators. New cohorts, measurements, perturbation results and longitudinal records can reveal systematic discrepancies, domain changes and processes missing from the current graph. Explicit opera- tor contracts help localise these gaps to a state, mapping, bridge or connection. Incorporating a new observation into a patient’s state posterior remains inference. Self-evolution occurs when accumulated evidence motivates a versioned change in the shared scientific model. Operator discovery through AutoResearch.Recent AI-enabled research systems show that agents can synthesise literature, generate hypotheses and propose experiments. They can also analyse biological data and refine hypotheses in light of new results[ 20,21]. In an AutoResearch workflow for Life Operators, agents could inspect residuals and identify missing variables. They could then propose equations or learned mappings, construct bridges and generate code. Such a workflow could move beyond fitting a fixed architecture and propose new Perception, Evolution or Generation operators when existing mappings fail. Each candidate should enter the proposal set with a scientific contract. The contract should 8 specify its inputs, outputs, assumptions, uncertainty, domain and failure conditions. Figure3 summarises this evidence-gated revision cycle. Its proposal and selection steps can be written schematically as C k+1 =Propose AI ( M k ,D dev ≤k ,∆ k ) , M k+1 =Select D ind k+1 (fM k g[C k+1 ),M k =(S k ,Ω k ,B k ,Γ Q,k ). (6) Here,S k denotes the state definitions, andΩ k denotes the library of Perception, Evolution and Generation operators. The symbolsB k andΓ Q,k denote the bridge-operator library and graph structure, respectively. Accumulated development dataD dev ≤k and discrepancies∆ k are used to generate a set of candidate model variantsC k+1 . A candidate may change any of these elements rather than merely adjust parameters. Selection by independent evidence.Independent data or discriminating experiments D ind k+1 are used to select among the current model and its proposed revisions. IncludingM k in the candidate set means that revision is not compulsory. Generating a proposal does not establish its validity. Each candidate should be compared with the current version and sim- pler alternatives using evidence matched to the intended claim. Held-out observations may support reconstruction or prediction, whereas intervention response may require perturba- tional, prospective or trial evidence. The same data should not both generate a candidate and provide its final confirmation. When evidence cannot distinguish candidates, the current model should be retained, its domain narrowed or structural uncertainty carried forward. Accepted operators enter a versioned library and can be recomposed for related medical questions. Repeated cycles of discrepancy detection, candidate generation and evidence- based selection allow the library and its graphs to change as data and knowledge accumulate. This is the intended sense of self-evolution: not unconstrained self-rewriting, but cumulative revision of an explicit, versioned scientific model. 5 Biomedical Applications With validated Operator Graphs that simulate intervention-conditioned biological trajecto- ries, medicine could move from describing disease to designing ways to change its course. Such a graph would provide an executable environment for designing and comparing can- didate interventions across biological scales. It could be revised as new data arrive. The central shift is from asking what is likely to happen to searching how a desired trajectory can be achieved. Personalised therapeutics.An Operator Graph could be used to design the therapeutic itself for a person, rather than only select from a fixed menu of standard treatments. Molecu- lar profiles, immune state, anatomy, physiology, exposures and prior responses could initialise an individual graph. Candidate molecules, biologics, cell or gene therapies, and combination regimens could then be generated and compared. The graph could also evaluate alternative doses, schedules and delivery strategies in terms of efficacy, toxicity, resistance and long- term outcomes. Personalised mRNA neoantigen cancer vaccines are one example, in which 9 mutation-derived neoantigens and HLA information guide an individual product[22,23]. The graph could then connect antigen presentation and immune dynamics to tumour response and recurrence. The same principle extends to other patient-specific drugs and therapeutic regimens, which could be redesigned as the patient’s state changes. Virtual development and trials.Operator Graphs could connect molecular target en- gagement, exposure, geometry or energy delivery to cellular response, organ function and clinical outcomes. Large candidate spaces for drugs, devices and procedures could be ex- plored across virtual cohorts representing different biological states. This would support the identification of response and toxicity subgroups, biomarkers, eligibility criteria, end- points and failure mechanisms. It would also help select the most informative laboratory experiments and clinical comparisons. In silico development would thus become an active design environment. It would focus physical experiments and trials on the candidates and populations most likely to be informative[24]. Adaptive precision care.For an individual patient, Perception operators would up- date the current state from imaging, physiological signals, biomarkers, omics, wearables and clinical records. Alternative drugs, doses, surgery, radiotherapy, ablation, device settings, behavioural interventions or watchful waiting could then be propagated through the same graph. Rather than returning only an endpoint risk, the model could compare intermedi- ate biological changes, time to benefit, adverse effects, recovery paths and future treatment options. Continued observations would reveal deviations from the expected trajectory and identify the next informative measurement. They would also allow prevention, treatment, rehabilitation and follow-up to be adjusted over time. Across patients, the resulting data would expose missing mechanisms and support the self-evolution of the operator library. These applications would not remove laboratory experiments or clinical trials. They would make them more targeted by identifying which candidates, patient groups and intervention contrasts are most worth testing. 6 Towards Life Modelling Infrastructure Realising these applications requires shared infrastructure and coordinated validation be- yond individual operators, task-specific graphs and institutions. It should match claims to evidence, test validity after composition, and support transparent reuse and revision across sites and populations. Evidence matched to claims.Life Operators should be evaluated against their intended claims. Observation reconstruction and uncertainty calibration require held-out measure- ments across relevant acquisition conditions. Interpretation of inferred parameters addition- ally requires identifiability and sensitivity analysis[ 25,26]. Temporal prediction should use prospectively separated trajectories at declared horizons and across sites and populations. Intervention-response claims require interventional evidence or observational evidence with an explicit causal identification strategy[ 27]. Clinical claims require evidence of benefit and safety in the intended workflow. Each operator and graph should declare its intended claim, supporting evidence and unresolved uncertainty. 10 Validity after composition.A valid operator does not automatically produce a valid Operator Graph. Measurement error, biological variation, parameter uncertainty and bridge design can interact during propagation and feedback. Validation should examine measurable intermediate states, semantic and unit consistency, and uncertainty transport. It should also assess intervention response, long-horizon stability and calibration. Graphs should be compared with simpler predictive, mechanistic or hybrid alternatives at comparable data and computational cost. An additional operator, scale or bridge is justified only when it improves the task, explains a failure or changes an experimental or clinical decision. Infrastructure for reuse and revision.Reuse requires machine-readable contracts and open, versioned registries for operators and bridges. Each entry should record state seman- tics, inputs and outputs, units and clocks, assumptions, and valid domains. It should also document provenance, evidence status, uncertainty and known failure conditions. Shared standards, reference implementations and benchmarks would allow the scientific and clinical community to reproduce components, compare alternatives and contribute new operators without rebuilding a graph. International collaboration for validation.No single institution can assemble dense multiscale and longitudinal measurements throughout the lifespan in diverse populations. Collaboration among population cohorts, hospitals, biobanks and registries can align data elements, intervention definitions and endpoints. When person-level data cannot be pooled, federated analysis can support joint evaluation across sites. Diverse cohorts are needed to test transportability across populations, healthcare systems, devices and clinical practice, and to reveal missing or non-transferable operators. Such collaborations should also sup- port independent testing of AI-proposed candidates and the maintenance of frozen clinical versions, audit trails and rollback procedures. A staged path.Development can begin with bounded Perception, Evolution and Genera- tion operators whose states can be measured independently. Bridges can then be validated using evidence at both scales. Minimal Operator Graphs can address questions with explicit interventions, measurable intermediate states and prospective endpoints. Shared registries and international cohort networks can provide benchmarks, external validation and evidence- gated reuse. This path begins with tractable questions rather than an attempt to construct a complete digital human. Each task adds tested operators, bridges and graph patterns to a shared system that expands with new data and knowledge. References [1]Xiaoxi Yao, David R. Rushlow, Jonathan W. Inselman, et al. Artificial intelligence- enabled electrocardiograms for identification of patients with low ejection fraction: a pragmatic, randomized clinical trial.Nature Medicine, 27:815–819, 2021. doi: 10.1038/ s41591-021-01335-4. [2]Bryan He, Alan C. Kwan, Jae Hyung Cho, et al. Blinded, randomized trial of sonographer versus AI cardiac function assessment.Nature, 616:520–524, 2023. doi: 10.1038/s41586-023-05947-3. 11 [3]Tao Tu, Mike Schaekermann, Anil Palepu, et al. Towards conversational diagnostic artificial intelligence.Nature, 642:442–450, 2025. doi: 10.1038/s41586-025-08866-7. [4]Valentin Liévin, Anil Palepu, Wei-Hung Weng, et al. Towards conversational ar- tificial intelligence for disease management.Nature, 655:1292–1299, 2026. doi: 10.1038/s41586-026-10764-5. [5]Artem Shmatko, Alexander Wolfgang Jung, Kumar Gaurav, Søren Brunak, Laust Hvas Mortensen, Ewan Birney, Tom Fitzgerald, and Moritz Gerstung. Learning the natural history of human disease with generative transformers.Nature, 647:248–256, 2025. doi: 10.1038/s41586-025-09529-3. [6]Sarah M. Urbut, Yi Ding, Tetsushi Nakao, et al. A Bayesian framework for lon- gitudinal EHR and genetic discovery.Nature, 656:924–935, 2026. doi: 10.1038/ s41586-026-10780-5. [7]Fei Liu, Kai Wang, Hui Xu, et al. Advancing cancer detection and treatment using longitudinal routine clinical data.Cell, 2026. doi: 10.1016/j.cell.2026.07.009. Online ahead of print. [8]Denis Agniel, Isaac S. Kohane, and Griffin M. Weber. Biases in electronic health record data due to processes within the healthcare system: retrospective observational study. BMJ, 361:k1479, 2018. doi: 10.1136/bmj.k1479. [9]Alan L. Hodgkin and Andrew F. Huxley. A quantitative description of membrane current and its application to conduction and excitation in nerve.The Journal of Physiology, 117(4):500–544, 1952. doi: 10.1113/jphysiol.1952.sp004764. [10]Alan M. Turing. The chemical basis of morphogenesis.Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences, 237(641):37–72, 1952. doi: 10.1098/rstb.1952.0012. [11]Hiroaki Kitano. Systems biology: a brief overview.Science, 295(5560):1662–1664, 2002. doi: 10.1126/science.1069492. [12]Jorge Corral-Acero, Francesca Margara, Maciej Marciniak, et al. The digital twin to enable the vision of precision cardiology.European Heart Journal, 41(48):4556–4564, 2020. doi: 10.1093/eurheartj/ehaa159. [13]Shuang Qian, Devran Ugurlu, et al. Developing cardiac digital twin populations powered by machine learning provides electrophysiological insights in conduction and repolarization.Nature Cardiovascular Research, 4(5):624–636, 2025. doi: 10.1038/ s44161-025-00650-0. [14]Lu Lu, Pengzhan Jin, Guofei Pang, Zhongqiang Zhang, and George Em Karni- adakis. Learning nonlinear operators via DeepONet based on the universal approx- imation theorem of operators.Nature Machine Intelligence, 3:218–229, 2021. doi: 10.1038/s42256-021-00302-5. 12 [15]Nikola Kovachki, Zongyi Li, Burigede Liu, Kamyar Azizzadenesheli, Kaushik Bhat- tacharya, Andrew Stuart, and Anima Anandkumar. Neural operator: learning maps between function spaces with applications to PDEs.Journal of Machine Learning Re- search, 24(89):1–97, 2023. URLhttps://w.jmlr.org/papers/v24/21-1524.html. [16]George Em Karniadakis, Ioannis G. Kevrekidis, Lu Lu, Paris Perdikaris, Sifan Wang, and Liu Yang. Physics-informed machine learning.Nature Reviews Physics, 3:422–440, 2021. doi: 10.1038/s42254-021-00314-5. [17]Francesca Margara, Yiangos Psaras, Zhinuo Jenny Wang, et al. Mechanism based ther- apies enable personalised treatment of hypertrophic cardiomyopathy.Scientific Reports, 12:22501, 2022. doi: 10.1038/s41598-022-26889-2. [18]Samira Merali, David H. Salinger, Maria Palmisano, et al. Recommendation of mava- camten posology by model-based analyses in adults with obstructive hypertrophic car- diomyopathy.CPT: Pharmacometrics & Systems Pharmacology, 13(9):1448–1461, 2024. doi: 10.1002/psp4.13138. [19]Iacopo Olivotto, Artur Oręziak, Roberto Barriales-Villa, et al. Mavacamten for treat- ment of symptomatic obstructive hypertrophic cardiomyopathy (EXPLORER-HCM): a randomised, double-blind, placebo-controlled, phase 3 trial.The Lancet, 396(10253): 759–769, 2020. doi: 10.1016/S0140-6736(20)31792-X. [20]Juraj Gottweis, Wei-Hung Weng, Alexander Daryin, et al. Accelerating scientific dis- covery with Co-Scientist.Nature, 655:487–496, 2026. doi: 10.1038/s41586-026-10644-y. [21]Ali E. Ghareeb, Benjamin Chang, Ludovico Mitchener, et al. A multi-agent sys- tem for automating scientific discovery.Nature, 655:497–505, 2026. doi: 10.1038/ s41586-026-10652-y. [22]Luis A. Rojas, Zachary Sethna, Kevin C. Soares, et al. Personalized RNA neoantigen vaccines stimulate T cells in pancreatic cancer.Nature, 618:144–150, 2023. doi: 10. 1038/s41586-023-06063-y. [23]Zachary Sethna, Pablo Guasp, Charlotte Reiche, et al. RNA neoantigen vaccines prime long-lived CD8+ T cells in pancreatic cancer.Nature, 639:1042–1051, 2025. doi: 10. 1038/s41586-024-08508-4. [24]Marco Viceconti, Francesco Pappalardo, Blanca Rodriguez, Marc Horner, Jeffrey E. Bischoff, and Flora Musuamba Tshinanu. In silico trials: verification, validation and uncertainty quantification of predictive models used in the regulatory evaluation of biomedical products.Methods, 185:120–127, 2021. doi: 10.1016/j.ymeth.2020.01.011. [25]Ryan N. Gutenkunst, Joshua J. Waterfall, Fergal P. Casey, Kevin S. Brown, Christo- pher R. Myers, and James P. Sethna. Universally sloppy parameter sensitivities in systems biology models.PLOS Computational Biology, 3(10):e189, 2007. doi: 10.1371/journal.pcbi.0030189. 13 [26]Alejandro F. Villaverde, Antonio Barreiro, and Antonis Papachristodoulou. Structural identifiability of dynamic systems biology models.PLOS Computational Biology, 12 (10):e1005153, 2016. doi: 10.1371/journal.pcbi.1005153. [27]Miguel A. Hernán and James M. Robins.Causal Inference: What If. Chapman & Hall/CRC, Boca Raton, 2020. URLhttps://miguelhernan.org/whatifbook. 14