{"id":"7a4fd39d-8967-4851-999f-195af89e2b7d","arxiv_id":"2507.08959","paper_version":1,"verdict":"REJECT","confidence":"LOW","novelty_score":3.0,"correctness_risk":"high","formal_verification":"none","parameter_count":9,"one_line_summary":"A hybrid GCN/GAT model with time windows reports AUC 0.937 on one private ad platform, but no baseline comparison supports the claimed improvement.","lead":"This paper applies a graph-based computer model to recommend ads across three platforms and reports a top score of 0.937 on one platform. The result could matter for ad targeting, but the paper does not compare against existing methods or release its data.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The central claim of 'improved' cross-platform ad recommendation is unsupported because the paper reports no baseline comparisons; absolute AUC on a private dataset cannot demonstrate enhancement.","rationale":"The reader's verdict REJECT is sound, but I would rest the rejection on a different primary pillar than the one listed as the weakest assumption. The reader's weakest_assumption focuses on the shared-user identity mapping and hand-assigned edge weight of 0.85. That is a real validity threat for the 'interest migration pathway' claim. However, the more fundamental and decisive problem is the absence of any baseline comparison: even if the identity mapping were perfect and the graph perfectly encoded true cross-platform behavior, the paper still provides no evidence that the GNN enhances recommendation over simpler alternatives. The reader's rationale does mention 'no baseline comparison' as one of several reasons, so there is partial agreement. The identity-mapping issue matters for the scientific interpretation of the pathways, but the missing comparators alone are sufficient to invalidate the headline claim of improved accuracy. I also note the garbled equations and the duplicated column header in Table 2 ('Accuracy (%)' appears twice, with Precision missing), which further reduce the ability to verify the model. My chosen concrete test (baseline comparison with statistical significance) would settle whether the claimed enhancement exists; if the GNN does not significantly beat a simple baseline, the central claim is unsupported. Therefore, I recommend leaving the reader's verdict unchanged.","tokens_in":5842,"tokens_out":4388,"duration_ms":56796,"concrete_test":"Re-run the method and at least three standard baselines on the same private dataset with the identical 70/15/15 split and hyperparameter search: (a) logistic regression on user/ad features, (b) a behavior-sequence model such as Deep Interest Network, and (c) LightGCN with the same graph but without time-aware cross-platform edges. Compute per-platform AUC, Precision, Recall, and F1 over five random seeds with mean and standard deviation. The 'enhancement' claim holds only if the GNN's AUC exceeds the strongest baseline by more than the combined standard error and a paired significance test (e.g., paired bootstrap or t-test) gives p < 0.05. If the gain is within noise, the central claim fails.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central claim is comparative: the abstract promises to 'improve the accuracy' of cross-platform advertisement recommendation, and the title says 'Enhanced.' Yet the experimental section provides no comparison against any baseline model. Table 2 lists only absolute metrics for the proposed GNN on the three platforms and the merged set; there is no row for logistic regression, matrix factorization, a non-graph neural recommender (e.g., Deep Interest Network), or even a GNN without time-aware edges. Without such a comparator, an AUC of 0.937 on an unshared 2,870-record dataset cannot establish that the graph architecture, attention layers, or cross-platform edges contribute to performance. Simple models often achieve high AUC on imbalanced click data, and the paper reports no error bars, confidence intervals, or significance tests. The phrase 'Platform B reaches 0.937 in AUC value, which is the best performance' likely means best among platforms, not best among algorithms, which makes the 'enhancement' claim even more ambiguous. The later 'comparative experiments' on weighted loss versus oversampling are reported as isolated numbers in prose without a table, seeds, or baseline comparison, so they do not remedy the gap. Even if the model equations were fully rendered (they are garbled as Eq. 1–3), a complete description would not supply the missing comparative evidence. Therefore, the load-bearing assertion that the method improves recommendation accuracy is not supportable from the submitted results.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a graph neural network (GNN) model for cross-platform advertisement recommendation, combining graph convolutional network (GCN) layers, graph attention network (GAT) layers, and time-aware encoding, with a shared-user-node strategy intended to capture cross-platform interest migration. The authors introduce a new edge type, 'view-cross-platform', connect the same user across platform nodes when semantically similar ads are clicked within 24 hours, and assign a hand-set edge weight of 0.85. Experiments are conducted on a private dataset of 2,870 records from three unnamed platforms, with a reported AUC of 0.937 on Platform B and 0.931 on the merged set. The paper claims that the method improves cross-platform advertisement recommendation accuracy and that hyperparameter tuning improves robustness on heterogeneous data.","tokens_in":6187,"tokens_out":6953,"duration_ms":83416,"significance":"If the comparative claim were established, the paper would offer a practical architecture for cross-platform ad ranking that integrates heterogeneous nodes (users, ads, platforms), temporal encoding, and cross-platform edges, with a plausible mechanism for modeling interest migration. The authors do provide useful descriptive details: hyperparameter ranges, preprocessing choices, evaluation metrics, and mitigation strategies for label imbalance. However, the current evidence is descriptive only. There are no baseline comparisons, no error bars or significance tests, no validation of the cross-platform identity mapping, and no public data or code. The central 'enhancement' claim is therefore unsupported as written, and the scientific value is limited to a method proposal with an unverified empirical narrative.","major_comments":[{"comment":"Table 2 reports metrics only for the proposed model on the three platforms and the merged set. The title and abstract claim that the method 'improves accuracy' and is 'enhanced', which are comparative statements. Without any baselines (e.g., logistic regression, matrix factorization, a non-graph deep recommender, or GNN ablations without time-aware or cross-platform edges), an absolute AUC of 0.937 on a private 2,870-record dataset cannot establish that the graph architecture contributes to performance. The phrase 'best performance' refers only to the best platform, not the best method. Please add baseline comparisons on identical splits, with repeated runs and significance tests.","section":"Experimental Results and Analysis, Table 2"},{"comment":"The evaluation tables list 'Accuracy (%)' twice and omit 'Precision (%)', although the text states that five metrics are used (AUC, Accuracy, F1, Precision, Recall). The first accuracy column (87.3, 89.1, 86.5, 88) matches the text's accuracy values, while the second (85.9, 88.2, 84.7, 86.6) appears to be precision, but this is never stated. The duplicated and mislabeled columns make the quantitative results impossible to interpret reliably, and the paper's numerical claims cannot be checked without a corrected table.","section":"Table 2 and Table 3, Model Training and Evaluation"},{"comment":"The mathematical presentation is incomplete and garbled. Equations (1), (2), and (3) contain placeholder or malformed symbols, and Equation (3) is used twice for two different formulas (the feature aggregation in Section 2.1 and the k-hop neighborhood expansion in Section 2.4). The GCN/GAT aggregation rule, the sequence recommendation scoring function, and the weighted cross-entropy loss cannot be verified from the text. Complete, correctly numbered equations with all variables defined are essential for a method paper.","section":"Graph Neural Network Model Architecture Design, Equations (1)-(3)"},{"comment":"The cross-platform edges are the central mechanism for the claimed 'interest migration pathways', but the construction rests on two unvalidated assumptions: that hashed identifiers plus temporal activity matching correctly align the same person across platforms, and that a hand-assigned edge weight of 0.85 for clicks on semantically similar ads within 24 hours is a meaningful representation of real interest migration. The paper reports no accuracy evaluation of the identity-matching step and no sensitivity analysis with respect to the 0.85 weight or the 24-hour window. Without such checks, the cross-platform edges may encode artifacts of the graph construction rather than measured user behavior.","section":"Shared User Node Strategy and View-Cross-Platform Edge, Model Construction"},{"comment":"All reported metrics come from a single 70/15/15 train/validation/test split, with no standard errors, confidence intervals, or multiple seed runs. Hyperparameters and class weights are tuned on the validation set and then used to report test performance, and the imbalance mitigation comparisons on Platform-C are reported as isolated numbers in prose. These limitations do not support the conclusion's robustness and adaptability claims. Please report variability across seeds or bootstrap resamples, and use nested or repeated validation for hyperparameter and class-weight selection.","section":"Experimental Protocol, Model Training and Evaluation"}],"minor_comments":[{"comment":"The text contains an unresolved 'Error! Reference source not found.' after the description of platform-node features, leaving the feature specification incomplete.","section":"Characterization of Cross-Platform Advertising Campaigns"},{"comment":"The figure references are inconsistent: Figure 1 is said to show temporal time windows, but later text refers to Fig. 1 for the architecture and edge types, while Figure 2 is also described as the GNN architecture. The figures should be renumbered and referenced in the correct order.","section":"Figures and captions"},{"comment":"In Table 1, the 'Sample of data inputs (10,000)' column reports a value of 42, which conflicts with the stated dataset size of 2,870 records; please clarify the units and values.","section":"Table 1, Hyperparameter Optimization"},{"comment":"The edge-type list is internally inconsistent: the text first lists edge types as 'view-platform', 'click-ad', and 'browse-user', then repeats the same names as edge attributes, and later mentions 'view-user' in reference to Fig. 1. Please clarify which names refer to edge types and which refer to edge attributes.","section":"Graph Edge Types and Attributes"},{"comment":"There are typographical and grammar errors throughout, including 'de-weighted', 'attention span' for 'attention heads', and the sentence 'Platform-C's recall and precision fluctuates greatly', which should agree in number. Also, the conclusion repeats the unsupported claim that the model 'effectively improves accuracy' without comparative evidence; the wording should match what the experiments actually demonstrate.","section":"Analysis of Experimental Results"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is far from publication in its current form, but the main omissions are potentially fixable. I recommend that if a revision is invited, the editor require baseline comparisons on identical splits, a corrected metric table, complete equations, validation of the cross-platform identity mapping with sensitivity analyses, and uncertainty quantification. I would be cautious about accepting the cross-platform interest-migration claim unless the identity-matching and edge-weight construction are independently validated; the current private dataset and the absence of code or data make external verification impossible."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a thin application paper that assembles GCN+GAT+TGAT with a shared user node for cross-platform ad recommendation, and the only empirical support is a single run on a private 2,870-record dataset with no baselines. The central word in the title and abstract – 'enhanced' – is never demonstrated.\n\nWhat's actually new: the shared-user-node strategy for cross-platform identity linking and a hand-designed 'view-cross-platform' edge type are reasonable engineering ideas for this problem. The authors also tried two class-imbalance fixes (weighted loss and oversampling) and report the effects in prose. Those are the only bits with any content.\n\nWhere it falls apart: the experiments. Table 2 lists only the proposed model's metrics per platform; there is no row for a non-graph baseline, no GNN without time-aware edges, nothing. An AUC of 0.937 on a private dataset tells you nothing about whether the graph structure or attention helps. The table also has a duplicated 'Accuracy (%)' header, so one of those columns is presumably Precision but mislabeled. Equations (1)–(3) are garbled in the PDF, and the 'comparative experiments' are described as isolated numbers without a table or any variance estimate. The dataset is tiny (2,870 records) yet the paper claims 'ultra-large scale' and talks about four A100s – that mismatch, plus no code/data, makes the whole thing unverifiable. The references are mostly marketing journals; no GNN recommendation systems literature is cited, so the authors don't engage with the actual state of the art they're building on.\n\nIs anything salvageable? The shared-user-node idea and the cross-platform edge definition could be a small component in a properly evaluated study. But as submitted, the paper is not a contribution; it's a descriptive report of one run.\n\nWho is this for? Possibly a practitioner who wants a high-level sketch of applying GNNs to cross-platform ads, but even that sketch is too vague to reproduce. No serious researcher should spend referee time on this in its current form. I'd desk reject with a message explaining that the enhancement claim needs baselines, a public dataset or benchmark, error bars, and corrected equations. If the authors resubmit with those, it could become a valid (if modest) application paper.","headline":"Thin application paper whose 'enhancement' claim is unsupported by any baseline comparison; the only new bits are a shared-user-node strategy and a cross-platform edge type, but the evaluation is one unverifiable run on a private dataset.","tokens_in":6650,"tokens_out":2560,"would_cite":false,"duration_ms":29876,"reading_group":"no","serious_thinker":"yes","would_accept_peer_review":false},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"By pooling users, ads, and platforms into one heterogeneous graph with shared user nodes and time-window edges, the paper's GNN model reaches AUC 0.937 on Platform B and claims to reveal how ad interests migrate across platforms.","keywords":["graph neural network","cross-platform advertisement recommendation","sequential recommendation","graph attention network","graph convolutional network","temporal encoding","interest migration","heterogeneous graph"],"falsifier":"Permute the user identities across platforms so 'view-cross-platform' edges connect different people, retrain on the same Platform-B data, and compare AUC: if the score remains near 0.937, the cross-platform edges carry no identity signal and the interest-migration claim is an artifact of graph construction.","tokens_in":5626,"feed_emoji":"🎯","tokens_out":8409,"duration_ms":96497,"temperature":0.7,"pith_summary":"This paper proposes that cross-platform ad recommendation can be improved by building a single graph whose nodes are users, ad spots, and platform devices, with edges for clicks, views, and a new type that links the same user across platforms when they click semantically similar ads within 24 hours. A hybrid GCN-GAT model with time-window encoding reaches an AUC of 0.937 on Platform B and 0.931 on the merged three-platform test set, the best numbers among the platforms; the authors attribute the gain to the shared user node letting the graph trace how interest migrates. The practical goal is that an advertiser running campaigns across platforms can use behavioral signals learned on one platform to rank ads better on another, rather than treating each platform's inventory independently. The authors also report that uneven ad-label distributions on Platforms A and C lower Precision and Recall, and that weighted loss or oversampling partially closes that gap.","feed_headline":"Cross-platform GNN ad recommender hits 0.937 AUC","feed_subtitle":"A shared-user graph learns interest migration paths; Platform B leads while imbalanced labels drag Platforms A and C down.","key_machinery":"The load-bearing object is a heterogeneous graph with three node types (users with 11 behavioral features, ad spots with 8 content features, platform devices with 6 environment features) and edge types 'view-platform', 'click-ad', and 'browse-user', plus the new 'view-cross-platform' edge that links the same user across platform nodes. The model is a stacked GCN (global neighborhood aggregation) and GAT (attention-weighted neighbor contributions), made time-aware by encoding user behavior into windows (Δt = 2h, 6h, 12h) via TGAT. The sequential recommendation score combines GAT attention weights with the user state vector and ad embedding through a feed-forward scoring function, and training uses weighted cross-entropy loss to counteract unbalanced ad labels. What this machinery is supposed to do is let the graph propagate interest signals from one platform to another through the shared user node rather than treating each platform's ad inventory as independent.","core_discovery":"On the paper's own terms, the discovery is that a heterogeneous graph built from user behavior, ad content, and platform features supports a trainable model—GCN for structure aggregation, GAT for attention-based neighbor weighting, and time-aware (TGAT-style) encoding for sequential behavior—that reaches AUC 0.937 on Platform B and 0.931 across the merged three-platform test set. The mechanism claimed to carry the improvement is the shared user node: the same hashed user appears once across platforms, and 'view-cross-platform' edges connect that node when the user clicks semantically similar ads within 24 hours, with edge weight 0.85 reflecting frequency and continuity. The paper reads the Platform-B result as evidence that these edges capture the latent pathways of user interest migration across platforms, and the per-platform differences as effects of label imbalance rather than architecture failure.","pith_inferences":["The paper never ablates the 'view-cross-platform' edges, so whether those edges actually carry the interest-migration signal is an open test, not a demonstrated fact.","The dataset has 2,870 records across three platforms; without confidence intervals or repeated-seed runs, the difference between Platform B's 0.937 and Platform C's 0.915 AUC may be sampling noise.","If ground-truth matched accounts were available, the 0.85 edge weight for similar-ad clicks within 24 hours could be estimated from data rather than hand-set, which would either confirm or undermine the migration-path story.","A natural extension is to replace the hashed-identity matching with privacy-preserving entity resolution; that would let the same graph construction work when raw user IDs cannot be shared across platforms."],"forward_implications":["On a platform with stable behavior sequences and cleaner labels, the model reaches AUC 0.937 (Platform B); on platforms with fragmented sequences and dense labels, AUC drops to 0.915–0.921.","The cross-platform merged evaluation reaches AUC 0.931 and F1 85.3%, which the paper reads as evidence the graph architecture copes with platform heterogeneity and ad diversity.","Handling label imbalance matters on Platforms A and C: weighted cross-entropy lifts F1 from 83.4% to 84.6% on Platform C, while oversampling raises Recall from 82.1% to 84.3% at a slight Precision cost.","Hyperparameters are platform-dependent: learning rate, batch size, and embedding dimension each shift training time, GPU occupancy, and stability, and batch size should be adjusted according to platform user active density."],"supporting_citations":[{"why":"cited as the basis for replacing GCN's fixed normalization with attention-computed neighbor weights in GAT.","marker":"[2]"},{"why":"cited as the basis for using Bayesian optimization on top of grid search to tune learning rate and reduce search cost.","marker":"[3]"},{"why":"cited as the basis for the sequential recommendation scoring function with GAT attention weights and user state vectors.","marker":"[5]"},{"why":"cited as the basis for the distributed training setup on four NVIDIA A100 GPUs.","marker":"[6]"},{"why":"cited as the basis for graph mini-batch inference with k-hop neighborhood expansion and memory control.","marker":"[7]"},{"why":"cited as the basis for the 70/15/15 train/validation/test split used in all experiments.","marker":"[9]"}],"fun_headline_variants":["GNN ad model tops 0.937 AUC on platform B","0.937 AUC: shared-user GNN maps ad interest drift","Time-aware GNN ad graph scores 0.937 AUC","Graph neural net ad recommender hits 0.937 AUC","Cross-platform GNN ad graph learns interest paths, 0.937 AUC"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that hashed IDs plus temporal activity matching correctly identify the same person across platforms and that a 0.85-weighted 'semantically similar ad within 24 hours' click is a genuine signal of interest migration; if either fails, the cross-platform edges encode artifacts rather than measured behavior.","fun_headline_variants_meta":{"raw":{"variants":["GNN ad model tops 0.937 AUC on platform B","0.937 AUC: shared-user GNN maps ad interest drift","Time-aware GNN ad graph scores 0.937 AUC","Graph neural net ad recommender hits 0.937 AUC","Cross-platform GNN ad graph learns interest paths, 0.937 AUC"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001515,"raw_usage":{"total_tokens":6039,"prompt_tokens":877,"completion_tokens":5162,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":493,"completion_tokens_details":{"reasoning_tokens":5071}},"tokens_in":493,"tokens_out":5162,"duration_ms":37379,"temperature":1.0,"reasoning_tokens":5071,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T18:08:36.545564+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Permute the user identities across platforms so 'view-cross-platform' edges connect different people, retrain on the same Platform-B data, and compare AUC: if the score remains near 0.937, the cross-platform edges carry no identity signal and the interest-migration claim is an artifact of graph construction.","supporting_citations":[{"cited_title":"The Rise of Influence Marketing in E-Commerce: a Review of Effectiveness and Best Practices [J]","cited_arxiv_id":null,"evidence_quote":"cited as the basis for replacing GCN's fixed normalization with attention-computed neighbor weights in GAT."},{"cited_title":"Research on Cross-Platform Digital Advertising User Behavior Analysis Framework Based on Federated Learning[J]","cited_arxiv_id":null,"evidence_quote":"cited as the basis for using Bayesian optimization on top of grid search to tune learning rate and reduce search cost."},{"cited_title":"The power of Sequence: A Qualitative Analysis of Consumer Targeting and Spillover Effects in Social Media Advertising [J]","cited_arxiv_id":null,"evidence_quote":"cited as the basis for the sequential recommendation scoring function with GAT attention weights and user state vectors."},{"cited_title":"Cross-Media Advertising Strategies and Brand Attitude: The Role of Cognitive Load [J]","cited_arxiv_id":null,"evidence_quote":"cited as the basis for the distributed training setup on four NVIDIA A100 GPUs."},{"cited_title":"Observing \"tuned\" advertising on digital platforms [J]","cited_arxiv_id":null,"evidence_quote":"cited as the basis for graph mini-batch inference with k-hop neighborhood expansion and memory control."},{"cited_title":"Mobile ad fraud: Empirical patterns in publisher and advertising campaign data [J]","cited_arxiv_id":null,"evidence_quote":"cited as the basis for the 70/15/15 train/validation/test split used in all experiments."}],"review_version":1}