FedTIP: Communication-Efficient Federated Temporal Prompting for Few-Shot Dynamic Graph Adaptation Authors: Xijun Wu, Xinming Zhang Published: 2026 ABSTRACT Temporal interaction streams are often distributed across data owners, while downstream labels and communication budgets remain limited. Existing temporal graph prompts typically adapt at a centralized trainer, whereas federated graph methods commonly exchange larger trainable states. This paper presents FedTIP, a prompt-level federated method for client-local few-shot adaptation of temporal interaction graphs. A centralized historical stage constructs a temporal knowledge bank whose encoder is frozen during downstream federation. Clients optimize interaction-conditioned prompts on local supervised events and transmit a compact prompt-side message. Prototype anchoring and a reliability rule based on support coverage and update magnitude combine heterogeneous client updates, while prompt-side proximal regularization and stage-scoped negative sampling preserve the temporal protocol. Experiments on Wikipedia, Reddit, and MOOC cover temporal node classification and transductive and inductive link prediction. FedTIP records the highest mean AUC-ROC in the nine reported task–dataset cells, with gains of 1.83–14.69 points over the strongest federated baseline in each cell. Its measured cumulative float32 uplink is 0.40 MiB per downstream task, 64.6–99.6% below the evaluated baselines. 1. Introduction Online services generate continuous streams of timestamped interactions among users, content, courses, and other entities. Temporal graph models preserve the order and timing of these events: JODIE and DyRep model evolving node states [1,2], whereas TGAT and TGN use temporal attention or memory-based message passing [3,4]. Their evaluation must also respect what was observable at each time. Surveys and benchmark studies identify chronological splitting, inductive status, and time-valid negative sampling as central concerns [5,6,7]. Thus, adapting a temporal graph model is constrained by both scarce downstream supervision and the history available before each prediction. In many deployments, organizations retain their own downstream evidence. Federated averaging learns from decentralized data through iterative aggregation of local model updates [8], and federated graph learning extends this access pattern to graph-structured data [9,10]. Communication can be reduced by jointly optimizing model compression and resource allocation [11] or by transmitting sparse gradients from heterogeneous devices [12]. Knowledge distillation offers another compact exchange mechanism for statistically heterogeneous clients [13]. These studies motivate the setting considered here: the client-owned observations form a chronologically evolving interaction stream, only a few labeled events are available for each downstream task, and repeated uplink transmission is constrained. Pre-training and prompting provide a natural starting point. Prompt tuning was introduced as a parameter-efficient way to adapt a frozen model [14]. Graph and dynamic graph pre-training learn reusable representations from historical structure [15,16,17]; GraphPrompt, GPF, and All-in-One then adapt graph encoders through compact prompts [18,19,20]. Temporal prompt methods introduce time-aware or node–time conditional adaptation [21,22], but their downstream support is normally available to a centralized trainer. Federated graph learning takes the complementary view that downstream observations remain distributed. Static or snapshot-based approaches exchange predictions or graph model parameters [23,24,25]. Federated dynamic-graph systems exchange temporal–spatial states, dynamic graph parameters, or task-specific temporal representations [26,27,28]; FedLink formulates federated temporal link prediction in a non-archival NeurIPS workshop manuscript [29]. These two lines of research leave a specific adaptation problem unresolved. Temporal prompting generally assumes centralized access to downstream support, whereas federated dynamic-graph methods coordinate substantially larger model or representation state. Simply applying FedAvg to a compact prompt does not resolve the resulting evidence imbalance: a fixed prompt cannot respond to event-specific partners, time intervals, and pre-event states, and a client with only a few chronological support events may observe one class or produce an update with an atypical magnitude. The challenge is therefore not only to reduce trainable parameters, but also to determine how event-conditioned updates and sparse class evidence should influence shared adaptation. Figure 1 illustrates the representation side of this problem. Continuous-time events differ in their pre-event node states, interaction partners, and time intervals, whereas a fixed task prompt applies one modulation direction to every event. An interaction-conditioned adapter can instead derive feature and temporal gates from each observable event context. On the aggregation side, client updates must also be interpreted in light of the sparse and potentially single-class support from which they were estimated. FedTIP addresses these coupled representation and aggregation problems by federating an interaction-conditioned adapter over a frozen temporal knowledge bank. Its key mechanism is a coverage-aware reliability rule that combines local label coverage with prompt update magnitude and applies the resulting weights consistently to both prompt deltas and class-anchor updates. Consequently, sparse client evidence influences the shared prompt and prototype geometry through the same reliability semantics. The structured uplink further admits a replace-one sensitivity bound for class prototypes and an additive sufficient-statistics decomposition for secure summation, while raw downstream support events remain with their owners. FedTIP builds on frozen temporal encoding, proximal regularization, prototypical objectives, and federated averaging; its methodological novelty lies in jointly governing event-conditioned prompt updates and class-anchor geometry through coverage-aware reliability under chronological few-shot support. The main contributions are summarized as follows:We formulate chronologically constrained federated temporal prompt adaptation and instantiate it with a frozen temporal knowledge bank and client-side interaction-conditioned adapters. Raw downstream support events remain with their owners, while recurrent federation exchanges structured prompt-side messages rather than the full temporal model.We introduce coverage-aware joint prompt–prototype aggregation. A reliability rule combines local label coverage and prompt update magnitude, and the resulting weights govern both prompt deltas and class-anchor contributions, coupling representation adaptation with prototype alignment under sparse and class-imbalanced client support.We characterize the structured message interface through a replace-one prototype-sensitivity bound and an additive sufficient-statistics decomposition, establishing compatibility with sensitivity-calibrated prototype release and secure summation.We evaluate FedTIP on three temporal interaction datasets and three task families. It obtains the highest reported mean AUC-ROC in all nine same-access cells among the evaluated methods; matched TGAT/TGN comparisons, component ablations, hyperparameter and partition analyses, and communication measurements isolate the behavior and cost of the proposed mechanism. We formulate chronologically constrained federated temporal prompt adaptation and instantiate it with a frozen temporal knowledge bank and client-side interaction-conditioned adapters. Raw downstream support events remain with their owners, while recurrent federation exchanges structured prompt-side messages rather than the full temporal model. We formulate chronologically constrained federated temporal prompt adaptation and instantiate it with a frozen temporal knowledge bank and client-side interaction-conditioned adapters. Raw downstream support events remain with their owners, while recurrent federation exchanges structured prompt-side messages rather than the full temporal model. We introduce coverage-aware joint prompt–prototype aggregation. A reliability rule combines local label coverage and prompt update magnitude, and the resulting weights govern both prompt deltas and class-anchor contributions, coupling representation adaptation with prototype alignment under sparse and class-imbalanced client support. We introduce coverage-aware joint prompt–prototype aggregation. A reliability rule combines local label coverage and prompt update magnitude, and the resulting weights govern both prompt deltas and class-anchor contributions, coupling representation adaptation with prototype alignment under sparse and class-imbalanced client support. We characterize the structured message interface through a replace-one prototype-sensitivity bound and an additive sufficient-statistics decomposition, establishing compatibility with sensitivity-calibrated prototype release and secure summation. We characterize the structured message interface through a replace-one prototype-sensitivity bound and an additive sufficient-statistics decomposition, establishing compatibility with sensitivity-calibrated prototype release and secure summation. We evaluate FedTIP on three temporal interaction datasets and three task families. It obtains the highest reported mean AUC-ROC in all nine same-access cells among the evaluated methods; matched TGAT/TGN comparisons, component ablations, hyperparameter and partition analyses, and communication measurements isolate the behavior and cost of the proposed mechanism. We evaluate FedTIP on three temporal interaction datasets and three task families. It obtains the highest reported mean AUC-ROC in all nine same-access cells among the evaluated methods; matched TGAT/TGN comparisons, component ablations, hyperparameter and partition analyses, and communication measurements isolate the behavior and cost of the proposed mechanism. The remainder of this paper is organized as follows. Section 2 reviews the related work. Section 3 presents the system model and problem formulation. Section 4 describes FedTIP. Section 5 presents the experimental setup, results, and limitations. Section 6 concludes the paper. 2. Related Work 2.1. Temporal Interaction Modeling and Prompt Adaptation Dynamic graph representation learning differs in how time is exposed to the encoder. Discrete-time methods transform an event stream into snapshots and apply static or recurrent graph encoders. GCN and GAT provide representative spatial operators [30,31], while ROLAND evolves node embeddings across snapshots [32]. Continuous-time DGNNs operate closer to the interaction stream. JODIE and DyRep model coupled trajectories or point-process dynamics [1,2]. TGAT and TGN use temporal attention or memory states [3,4], while TREND and CAW capture event-node dynamics or causal walks [33,34]. Surveys establish the need for chronological and inductive evaluation [5]; benchmark studies specify time-valid splits and negative-sampling protocols [6,7]. Pre-training reduces dependence on downstream labels by learning reusable graph representations. DDGCL and CPDG exploit temporal regularities in historical streams [16,17]. Prompt learning changes the unit of adaptation: GraphPrompt, GPF, and All-in-One use compact graph prompts [18,19,20], and surveys organize these methods as parameter-efficient graph adaptation [35,36]. Temporal prompting introduces time-aware or node–time conditional prompts [21,22]. 2.2. Communication-Efficient Federated Adaptation FedAvg establishes the iterative model-averaging baseline for decentralized learning [8]. Communication-aware variants reduce transmitted states through model compression [11], sparse gradients [12], or distillation [13]. Statistical heterogeneity motivates changes to aggregation. FedProx constrains local drift, SCAFFOLD reduces client update variance, and FedNova corrects objective inconsistency under unequal local computation [37,38,39]. FedDyn adds dynamic regularization, while FedBN retains local normalization statistics under feature shift [40,41]. FedProto establishes a prototype-level federation pattern in which clients upload class prototypes, the server aggregates them, and global prototypes return to clients [42]. FedHide subsequently characterizes leakage from shared class prototypes and replaces true prototypes with proxy representations [43]. Differentially private prototype learning further shows how a frozen public encoder and private class representations can be coupled through sensitivity-calibrated prototype release [44]. Ditto relates regularized local adaptation to personalized federation [45]. 2.3. Federated Graph and Dynamic Graph Learning Federated graph learning makes graph ownership and partitioning explicit. Surveys organize the field around data privacy, graph heterogeneity, and structured collaboration [9,10]. FedGNN addresses privacy-preserving recommendations [23], FedSage+ generates missing cross-client neighbors [24], and FedGCN studies communication–convergence trade-offs for graph convolution [25]. FedGTA uses topology-aware aggregation [46]; FedGraphNN and FedGraph provide benchmark infrastructure [47,48]. Federated dynamic graph learning provides the nearest temporal baseline family. Feddy and STFL study dynamic GNN federation or temporal–spatial graph learning [26,27]. FedDGL models heterogeneous temporal evolution [28], and FedLink addresses federated dynamic link prediction [29]. Taken together, prior studies address temporal prompting, communication-efficient federation, prototype aggregation, and federated dynamic graph learning largely as separate problems. They do not jointly model event-conditioned prompt adaptation and class-prototype aggregation when chronological supervision is both client-owned and extremely sparse. 3. System Model and Problem Formulation FedTIP considers a coordinating server and independently owned clients under a four-part access model. Historical interactions are available centrally for pre-training; sampled downstream support events are client-owned and used for local adaptation; a global downstream validation segment is visible to the server for model selection; and a global test segment is visible to the server only for final scoring. The locality scope covers the few-shot downstream support sets within this centralized pre-training and evaluation protocol. The objective is to adapt a temporal interaction model while keeping those support events with their owners, respecting chronological visibility, and measuring cumulative uplink payload. 3.1. Temporal Interaction Graph Let G=(V,E) be a temporal interaction graph represented as a chronologically ordered event stream(1)E={en=(un,in,tn,yn,xn)}n=1N,t1≤t2≤⋯≤tN. Here un is a source-side node, in is a destination-side node, tn is the interaction timestamp, yn is an optional dynamic state label associated with the event or source node, and xn∈Rde is an edge feature vector. For any time t, Ht− denotes the events observable strictly before t. Before training, the manifest assigns every raw event to one of six mutually exclusive outcomes: historical pre-training, pre-training validation, an observed downstream supervised pool Esup, an inductive-node exclusion from the downstream training candidate window, downstream validation Eval, or downstream test Etest. The inductive test set is a flagged subset of Etest involving a typed user or item node removed from Esup. Earlier client updates and training negatives draw exclusively from information available before validation and test. The global downstream validation segment selects method-specific hyperparameters and is substantially larger than the 30-event support set; this stronger server-visible resource is part of the reported protocol. 3.2. Deployment and Communication Model The server coordinates K clients and maintains the shared model state. Before the first downstream task, it deploys the frozen temporal bank to participating clients. In round r of task τ, it broadcasts the current adapter and anchor prototypes and receives the FedTIP uplink message (2)Mk(r)=ΔΩk(r),{qk,c(r),nk,c}c∈Yk,bk,∥ΔΩk(r)∥2. The first component is the prompt-adapter delta; the remaining components are class prototype means and counts, a label-coverage score, and the update norm. This tuple constitutes the FedTIP uplink message. Raw downstream support events remain with their assigned clients, while the centrally constructed frozen temporal encoder is deployed separately before downstream adaptation. The system abstraction assumes that every selected active client completes its local computation and upload within a round. Communication accounting sums float32 values transmitted from clients to the server over all active clients and rounds for one downstream task. Cumulative uplink payload provides the common cross-method measurement, while Section 5.4 analyzes initialization and recurrent downlink from each method’s exchanged objects and communication cadence. 3.3. Federated Few-Shot Adaptation There are K clients C={1,…,K}. In the main user-iid profile, the split manifest first fixes a user-ownership map πU:VU→C and then assigns every downstream event e=(u,i,t,y,x) to π(e)=πU(u). Consequently, all events of one user have exactly one client owner. Client k owns a supervised pool Esup,k, and the client pools form a disjoint union of Esup. Label-skewed and activity-skewed profiles change the user-to-client map but retain the same pre-training-time materialization and method-independent ownership rule. For a downstream task τ, an active client set Aτ⊆C is selected from the manifest. Each active client receives a small local support set(3)Dτ,k⊆Esup,k. The support rows and their labels remain inside the assigned client boundary. The manifest also assigns validation and test rows to their user owners; the disjoint client views form the corresponding global segments, whereas model selection and final scoring use their unions at the server. This definition is method-agnostic: a baseline may adapt a prediction head, a prompt, or a larger model state, but every method follows the same support ownership, active-client set, chronological split, and common server-side evaluation segments. 3.4. Evaluation Tasks The evaluation contains three task families. Temporal node classification predicts a state label for the source-side node at the interaction time. Transductive link prediction evaluates later interactions among nodes observed before the downstream test stage. Inductive link prediction restricts scoring to test events containing at least one typed node excluded from the downstream supervised pool. Each positive link example is paired with one negative destination. A negative event (u,j,t) is valid only if the interaction is absent at time t under the stage-specific history allowed by the protocol. All methods use the same deterministic stage-valid sampling rule and task/split seed. We materialize this rule for all 9000 task-support events, and Section 5.1 reports the resulting collision checks. 3.5. Trust and Privacy Boundary Attack-oriented multimedia-security studies evaluate exposed intermediate structures under explicit adversarial capabilities; for example, video cryptanalysis considers chosen-plaintext, known-plaintext, and chosen-ciphertext attacks against internal encryption transformations [49]. Following this principle, FedTIP’s threat model specifies message visibility and manipulation surfaces for prompt and prototype disclosures. The protocol assumes that the coordinator and clients execute the prescribed optimization and reporting rules. The confidentiality analysis treats the coordinator as honest-but-curious after it receives the structured uplink messages, and the integrity analysis treats prompt deltas, prototype summaries, counts, coverage values, and update norms as possible manipulation surfaces. Raw downstream support events remain with their manifest-assigned owners. The coordinator observes Mk(r); a participating client observes the frozen bank, the current adapter, and server-aggregated anchor prototypes. Client-specific prototypes are used to construct the global anchors and are not redistributed to other clients. The bank broadcast defines the reverse disclosure path because its centrally derived historical statistics and negative-sampling support are visible to participating clients. The prototype message admits a direct sensitivity analysis. Let the representations used in class c be clipped so that ∥h˜k,e∥2≤Cemb, and consider fixed-size replace-one adjacency for Dk,c. Since (4)qk,c=1nk,c∑e∈Dk,ch˜k,e, its l2-sensitivity is bounded by (5)Δ2(qk,c)≤2Cembnk,c. The bound decreases with the local class count. For nk,c=1, the prototype equals the clipped representation of that event, making the small-count case the most informative message. This result formalizes the prototype exposure identified in prototype-sharing studies [43]. Across the 500 task–client pairs per dataset, the realized supports form 604, 602, and 625 local class prototypes on Wikipedia, Reddit, and MOOC. Of these, 104 (17.22%), 102 (16.94%), and 118 (18.88%) are single-event prototypes. These counts characterize how often the sensitivity bound operates in its highest-exposure small-support regime. FedTIP’s weighted prompt and anchor updates can be expressed through four additive sufficient statistics:(6)SΩ=∑k∈AτnkrkΔΩk,ZΩ=∑k∈Aτnkrk,Sc=∑k:c∈Yknk,crkqk,c,Zc=∑k:c∈Yknk,crk. The server updates use SΩ/ZΩ and Sc/Zc; therefore, the aggregation rule is compatible with secure summation of the numerators and denominators [50] without changing the client objective or prompt parameterization. In the main profile, these statistics combine the five active clients selected for each task, and clients receive the resulting global adapter and anchors. Equation (5) also gives a differential-privacy-compatible prototype map(7)q˜k,c=qk,c+N0,σ2Δ22I, where Δ2=2Cemb/nk,c links the perturbation scale to the proved prototype sensitivity. This construction connects the FedTIP message interface to established differential-privacy mechanisms and private prototype learning [44,51]. 4. FedTIP Method FedTIP federates an interaction-conditioned prompt adapter while keeping the temporal encoder frozen. Its central mechanism is coverage-aware joint prompt–prototype aggregation: reliability derived from local label coverage and update magnitude weights both prompt deltas and class-anchor contributions. A temporal knowledge bank supplies historical representations, while the shared reliability semantics govern how heterogeneous few-shot evidence changes the global adapter and prototype geometry. 4.1. Overview and Federated Visibility Figure 2 organizes FedTIP into three stages. Stage I constructs the frozen bank B0 from chronological historical interactions. In Stage II, parallel feature and temporal branches map each client-visible event context to an interaction-conditioned representation. In Stage III, clients transmit Mk(r) from Section 3; the server updates the adapter and anchor prototypes. The encoder, temporal memory, neighbor states, and time encodings are frozen after the centralized historical stage. Downstream visibility and communication follow Section 3; repeated client-to-server transmission is confined to Mk(r). Five elements define the method. The bank B0 contains the frozen temporal encoder and historical metadata. The adapter Ω contains the feature gates, temporal gates, and task template. Local prototypes {qk,c} summarize class geometry, while global anchor prototypes are maintained separately at the server. The reliability score rk uses label coverage and update magnitude. The filtered negative index Ineg is a protocol object, separate from the neural modules. The bank precedes downstream task sampling, local updates use the frozen bank and preceding events in the assigned task support, and server-side adaptation uses only the FedTIP uplink messages. 4.2. Temporal Knowledge Bank and Interaction Prompting The temporal knowledge bank defines the shared information available before downstream labels are sampled. It is broader than a parameter checkpoint because it also contains time encodings, historical statistics, and negative-sampling support for permitted stages. The prompt adapter then specializes this shared temporal knowledge to a client-local downstream event without changing the frozen encoder. Anchor prototypes are initialized separately on the server and evolve only during downstream collaboration. During historical bank construction, the frozen encoder produces temporal node states(8)hut−,hit−=fΘ0(u,i,t,Ht−), where Ht− denotes the chronological history before t in the pre-training stream. The encoder is trained with temporal link prediction on the historical stream. Given a positive interaction (u,i,t), a negative destination is sampled from the pre-training filtered candidate set(9)Ntpre(u)={j∈VI:(u,j,t)∉Hpret}, where VI is the destination-side node set and Hpret contains pre-training interactions observable up to time t. The filtering rule prevents a same-time or historical positive interaction from being used as a negative example whenever the protocol can detect it. The pre-training objective is (10)Lpre=−logexp(s(hut−,hit−)/τ)exp(s(hut−,hit−)/τ)+exp(s(hut−,hjt−)/τ), where j∼Ntpre(u), s(·,·) is cosine similarity or a bilinear score, and τ is a temperature. The output of this stage is the frozen bank(11)B0={fΘ0,TE,Shist,Ineg}, which contains the encoder, time encoding, historical recency and frequency statistics, and stage-scoped negative-sampling support. The bank provides temporal knowledge to downstream clients without requiring them to learn long-range event dynamics from a handful of labels. During downstream adaptation, client-side state construction uses the frozen historical bank and the client’s task support observed before the current event. For client k in task τ, the event state is written as(12)hk,ut−,hk,it−=fΘ0(u,i,t,Hpret−∪Hτ,kt−), where Hpret− is the centrally constructed bank context and Hτ,kt−={e∈Dτ,k:tedFedTIP↓ denotes the corresponding per-task recurrent costs, the symbolic amortization point is Tm⋆=⌈Ninit|B0|/(dm↓−dFedTIP↓)⌉. 5.5. Ablation Study Table 4 reports absolute AUC-ROC values for the full model, five single-component variants, and a joint prototype–aggregation variant. Each value first averages the 100 generated tasks within a dataset and then assigns equal weight to Wikipedia, Reddit, and MOOC. A cross denotes a removed component, whereas a blank entry denotes that the component is retained. The variants preserve the main protocol and change only the stated mechanism. Removing TKB keeps the encoder architecture and downstream interface but omits historical self-supervised pre-training before the encoder is frozen. Removing ICP replaces the interaction-conditioned prompt with a static shared prompt. Removing PA sets λa=0 and suppresses global-anchor exchange and updates while retaining the local class prototypes required by node classification. Removing RWA sets rk=1, which gives support-weighted FedAvg, wk=nk/∑jnj, and class-anchor weights ak,c=nk,c. Removing FTN uses unfiltered destination sampling in both pre-training and downstream adaptation while still excluding the current positive destination. Variant 6 removes PA and RWA together and keeps TKB, ICP, FTN, and all remaining settings fixed. As shown in Table 4, removing TKB decreases NC, T-LP, and I-LP by 5.11, 7.99, and 8.18 points relative to full FedTIP, the largest losses in the table. Removing ICP decreases the three task-family means by 3.18, 4.88, and 5.49 points. These patterns connect historical temporal knowledge and event-conditioned prompting to adaptation from sparse supervision. FTN is more task-specific: its removal decreases NC by 0.78 points but T-LP and I-LP by 5.54 and 5.73 points, respectively, consistent with negative destinations entering the link objective directly. Table 4 further shows that removing PA decreases the three means by 2.52, 4.42, and 4.58 points, while replacing RWA with support-weighted FedAvg decreases them by 2.86, 4.12, and 4.40 points. Their joint removal produces larger losses than either single removal—3.50, 5.32, and 6.05 points—but smaller losses than the sum of the two single-component effects. To quantify this pattern, for task family τ we define(35)IPA,RWAτ=ΔPA+RWAτ−ΔPAτ−ΔRWAτ, where ΔXτ=Mfullτ−M−Xτ. The resulting interaction contrasts are −1.88, −3.22, and −2.93 for NC, T-LP, and I-LP. The negative contrasts describe sub-additive overlap between PA and RWA, consistent with both mechanisms acting through the server-side prototype and client update aggregation path. 5.6. Core Hyperparameter Sensitivity Figure 4 examines the three coefficients that directly regulate downstream adaptation. Each one-at-a-time sweep uses the same 100 Wikipedia task identifiers under the main low-supervision setting, with the other two coefficients and all protocol settings fixed at their main values. The error bars quantify task-level dispersion across these controlled comparisons. Figure 4 shows that the selected λp=10−2 improves NC, T-LP, and I-LP by 1.44, 1.08, and 1.08 points over λp=0, and by 2.52, 2.02, and 2.05 points over λp=1. This interior optimum balances client-specific adaptation against proximal control: removing the term permits greater local drift, whereas an excessive coefficient suppresses useful prompt updates. Similarly, λa=10−1 improves the three tasks by 2.64, 2.73, and 2.65 points over no prototype anchoring and by 1.67, 1.19, and 1.27 points over λa=1, indicating that moderate class-level alignment is useful without forcing heterogeneous clients too tightly toward the global anchors. For anchor momentum, β=0.1 exceeds β=0 by 1.75, 1.52, and 1.64 points and β=1 by 1.89, 1.64, and 1.85 points on NC, T-LP, and I-LP. The result favors strong historical smoothing with a modest current-round update: β=0 completely preserves the previous anchor and prevents incorporation of current client evidence, whereas β=1 follows the current-round aggregate without historical carry-over. Across all three sweeps and task families, the selected values attain both the highest mean and the smallest observed task-level standard deviation. 5.7. Client Coverage, Support Scaling, and Partition Skew Figure 5 varies the active-client count A from two to ten on Wikipedia and sets total support to 6A events. The horizontal trend therefore measures the joint effect of broader client coverage and increased supervision. At fixed A, user-iid, label-skew, and activity-skew use the same support count, isolating the effect of the partition profile. As shown in Figure 5, under user-iid partitioning, increasing A from two to ten raises AUC-ROC from 72.24 to 73.36 for node classification, 92.22 to 93.27 for transductive link prediction, and 90.10 to 91.24 for inductive link prediction. The five-client points (72.91, 92.86, and 90.81) match the Wikipedia FedTIP row in Table 2. Beyond five clients, each curve gains less than 0.5 points. Figure 5 further shows that, at five active clients, label skew is 1.47, 0.83, and 0.93 points below user-iid, while activity skew is 1.28, 1.12, and 1.19 points below. At ten clients, the respective gaps are 1.52, 0.89, and 1.01 points for label skew and 1.31, 1.19, and 1.28 for activity skew. 5.8. Limitations and Scope The study has three principal scope boundaries. First, historical pre-training and downstream validation are centralized; the locality scope covers downstream support events. Second, the cross-method quantitative bars cover cumulative float32 uplink under complete participation of the selected clients. Downlink is characterized from each method’s exchanged objects and communication cadence: FedTIP combines one-time frozen-bank deployment with recurrent adapter-and-anchor broadcasts, whereas the compared full-state methods repeatedly distribute substantially larger model or auxiliary state. Protocol headers and retransmissions remain outside this parameter-payload analysis. Third, the structured uplink exposes prompt updates, prototype summaries, counts, and reliability fields to the coordinator. Section 3.5 characterizes this message-level threat surface, derives the small-count prototype sensitivity, and establishes secure-summation and sensitivity-calibrated interfaces for the aggregation rule. 6. Conclusions FedTIP formulates the temporal prompt adapter as the federated trainable object and uses coverage-aware joint prompt–prototype aggregation to govern both prompt deltas and class-anchor geometry. On Wikipedia, Reddit, and MOOC, it has the highest reported mean AUC-ROC in the nine same-access cells and a measured cumulative uplink of 0.40 MiB per task. The message-level analysis further characterizes prototype exposure and shows how FedTIP’s additive aggregation interface composes with secure summation and sensitivity-calibrated prototype release. Together, these results establish the practical value of prompt-level federation for chronologically ordered few-shot support.