{"id":"66902186-13d1-4b9c-87d2-115dc51f0654","arxiv_id":"2607.06349","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":6,"one_line_summary":"An LLM-based agentic sampler over building knowledge graphs selects target-specific exogenous variables for zero-shot IoT forecasting, matching or beating trained baselines on three real buildings.","lead":"TopoBrick uses an AI agent to read building floor-plan graphs and pick which other sensors, weather data, and schedules should help forecast each target sensor—without training on the target building. It beats zero-shot baselines and matches trained models on three real buildings, though most gains come from future-known weather data rather than the topology-selected sensor variables themselves.","discovery_kind":"unclear","skeptic_critique":{"model":"glm-5.2","headline":"The ablation in §4.4 shows the agentic topology sampler's unique output (past-known sensor selection) provides minimal or negative benefit, while gains come from weather/schedule injection that doesn't require topology-aware selection—undermining the central contribution.","rationale":"The reader correctly identified the most important tension in the paper. The ablation evidence is internally consistent and clearly shows that the agentic topology sampler's primary output—target-specific past-known sensor selection—does not improve forecasts. The gains come from future-known variables (weather forecasts, schedules) that are building-level and do not require topology-aware reasoning to select. This creates a gap between the paper's framing (topology-aware sampling as the key innovation) and the empirical evidence (weather/schedule injection as the actual driver of gains). The sampler comparison in §4.3 partially addresses this by showing the agentic sampler outperforms alternatives, but it does not decompose by variable type, so it cannot distinguish whether the sampler's advantage comes from past-known or future-known selection. The concern is concrete and testable. I recommend keeping the verdict at CONDITIONAL because the paper's practical contribution (availability-aware formulation + weather injection for zero-shot building forecasting) may still be valuable even if the agentic sampler itself is not the load-bearing component. However, the paper's central claim as stated—that topology-aware exogenous-variable selection provides a stronger inductive bias—is not fully supported by the presented evidence. The proposed decomposition test would settle whether the sampler contributes beyond what simple inclusion of all weather and schedule variables provides. If the sampler's advantage disappears under decomposition, the paper should reframe its contribution around the availability-aware formulation rather than the agentic topology sampler. No statistical significance tests or error bars are reported, which compounds the uncertainty, but this is secondary to the attribution problem.","tokens_in":20587,"tokens_out":2921,"duration_ms":265370,"concrete_test":"Run the sampler comparison from §4.3 (Figure 2) under two decomposed conditions: (a) each sampler selects ONLY past-known sensor variables, with calendar+schedules+oracle weather fixed identically across all methods; (b) each sampler selects ONLY future-known variables, with no past-known sensors. If in condition (a) the agentic sampler's advantage over random/ontology/k-hop is negligible (consistent with the §4.4 ablation showing +PastObs doesn't help), and in condition (b) the advantage is also negligible (because weather and schedules are building-level and don't require target-specific topology), then the agentic topology sampler is not responsible for TopoBrick's gains. Additionally, compare full TopoBrick against a simpler variant that uses ALL weather + ALL schedules + calendar (no sampler, no past-known sensors); if this variant matches TopoBrick within 2% nMAE, the sampler's con","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The paper's central claim is that topology-aware exogenous-variable selection via the agentic sampler provides a stronger inductive bias than alternatives. However, the ablation in §4.4 decomposes the contribution by covariate group and reveals that +PastObs—the primary output of the agentic topology sampler—provides minimal or negative benefit (+4.7% nMAE on BTS-B at H=24, +10.0% at H=48). The gains come from +WeatherFcst (−8.5% on LBNL59, −6.6% on BTS-B) and +OpSched (up to −7.3%). Critically, weather variables and operational schedules are building-level global variables (§3.1.2, §3.2.1) that do not require target-specific topology reasoning to select—one simply includes all weather channels and all schedules. The sampler comparison in §4.3 (Figure 2) shows the agentic sampler outperforming random/ontology/k-hop alternatives, but this experiment uses oracle weather for all methods and does not decompose whether the sampler's advantage comes from selecting better past-known sensors or better future-known variables. Since the ablation shows past-known sensor selection doesn't help, and future-known variables (weather, schedules) don't require topology-aware selection, the agentic topology sampler—the paper's central novel contribution—may not be the actual source of TopoBrick's performance gains. The gains may instead come from the availability-aware formulation (a design choice, not a contribution of the sampler) and weather forecast injection (a standard covariate, not topology-dependent).","agreement_with_reader":"agree"},"referee_report":{"model":"glm-5.2","summary":"TopoBrick proposes a training-free framework for zero-shot building IoT forecasting that uses an LLM-based agentic topology sampler over building knowledge graphs (Brick schema) to select target-specific exogenous variables. The selected variables are organized by deployment-time availability into past-known (sensor states, masked over the prediction horizon) and future-known (calendar, operational schedules, meteorological forecasts) groups, then fed to a frozen time-series foundation model. The framework is evaluated on three real-world buildings (LBNL59, BTS-B, BTS-C) across four prediction horizons, against naive baselines, fully trained building-specific models, and zero-shot foundation models (Chronos-2, Moirai, TimesFM). The main results show TopoBrick achieving best or second-best performance in the majority of settings, with ablations decomposing contributions by covariate group and sampler strategy.","tokens_in":21374,"tokens_out":1326,"duration_ms":436771,"significance":"The paper addresses a genuine gap: existing zero-shot time-series foundation models lack mechanisms for selecting which building points serve as exogenous variables, and building-specific supervised models do not scale across heterogeneous portfolios. The idea of using building KG topology as a routing layer for exogenous variable selection is well-motivated. Strengths include reproducible code (GitHub link provided), evaluation on three real buildings across two continents, a deployment-faithful past/future-known variable split, and a falsifiable ablation design that honestly reports where the method helps and where it does not. The per-ontology analysis (Table 3) is a valuable contribution that connects performance to physical interpretability.","major_comments":[{"comment":"§4.4, Figure 3: The ablation reveals that +PastObs—the primary output of the agentic topology sampler—provides minimal or negative benefit (+4.7% nMAE on BTS-B at H=24, +10.0% at H=48), while gains come primarily from +WeatherFcst and +OpSched. Since weather and schedules are building-level global variables that do not require target-specific topology reasoning to select, this raises a question about whether the agentic topology sampler—the paper's central novel contribution—is actually responsible for the performance gains. The paper should either (a) run an ablation where weather and schedules are added without the agentic sampler (e.g., using all available past-known sensors selected by random or k-hop) to isolate the sampler's contribution, or (b) explicitly reframe the central claim to acknowledge that the availability-aware formulation and weather injection are the primary drivers,","section":null},{"comment":"§4.3, Figure 2: The sampler comparison uses oracle weather for all methods but does not decompose whether the agentic sampler's advantage over random/ontology/k-hop comes from selecting better past-known sensors or better future-known variables. Given that §4.4 shows past-known sensors provide minimal benefit, the sampler comparison may conflate the sampler's variable-selection quality with the universal benefit of weather. A decomposition of the sampler comparison by covariate group (past-known vs. future-known) would clarify whether the agentic sampler's advantage is actually in topology-aware sensor selection or simply in consistently including weather/schedule variables that simpler alternatives might miss. This is load-bearing for the claim that topology-aware sampling is more reliable than simpler alternatives.","section":null},{"comment":"Table 2: No error bars, confidence intervals, or statistical significance tests are reported for any result. Several key comparisons involve small margins (e.g., TopoBrick nMAE 0.319 vs. PatchTST 0.319 on BTS-C at H=24; TopoBrick nMAE 0.476 vs. PatchTST 0.473 on BTS-C at H=72). Without variance estimates, it is unclear whether these differences are meaningful. At minimum, bootstrap confidence intervals or paired tests across the per-sensor results should be reported for the key comparisons.","section":null}],"minor_comments":[{"comment":"§3.1.3: The LLM model choice (gpt-oss-20b) is mentioned only in §4 settings. A brief note in the methodology about model choice and sensitivity would help readers assess generalizability, or alternatively a statement that the approach is model-agnostic.","section":null},{"comment":"Table 3: The per-ontology results are extensive but the color coding (green/red) is not visible in the text rendering. Consider adding sign indicators (+/−) or ensuring the color encoding is accessible in both print and digital formats.","section":null},{"comment":"§4.4, Figure 3: The BTS-C ablation panels (f, i, l) appear to have fewer bars than other buildings, and the text notes BTS-C 'does not seem to respond well to our ablation modules.' This is mentioned briefly but deserves a clearer explanation or cross-reference to §5 discussion.","section":null},{"comment":"The self-citation to Lin et al. 2024 (Bitsa, Ref [6]) is relevant but should be clearly distinguished from the current contribution to avoid confusion about novelty. A sentence clarifying what is new versus what was established in prior work would help.","section":null},{"comment":"§3.1.1: The skeleton construction removes Point leaves from the traversal graph, but the notation V_s = {u ∈ V ∖ V_pt | τ(u) ∈ {Equipment, Location}} excludes other structural node types (e.g., System). The text mentions 'systems' as structural nodes but the formal definition does not include them. Clarify whether System nodes are included or mapped to Equipment/Location.","section":null},{"comment":"The ACM template artifacts (conference acronym, copyright year 2018, Woodstock NY) suggest this is a draft submission. Ensure formatting is corrected for the final version.","section":null}],"recommendation":"major_revision","confidential_remarks":"The stress-test concern about the ablation undermining the central contribution is well-founded and is the primary reason for the major_revision recommendation. The paper's honesty in reporting the +PastObs ablation results is commendable, but the framing of the central claim needs to be reconciled with the evidence. If the authors can show that the agentic sampler's advantage persists even when weather/schedules are held constant across methods (which Figure 2 partially does but does not fully decompose), the contribution would be substantially strengthened. Alternatively, reframing the contribution as 'availability-aware exogenous variable formulation with topology-guided selection' rather than 'topology-aware sampling as the primary driver' would also address the concern."},"author_rebuttal":{"model":"glm-5.2","summary":"We thank the referee for a thorough and constructive report. The referee raises three major points: (1) whether the agentic topology sampler—the paper's central contribution—actually drives performance gains, given that the ablation shows past-known sensors (the sampler's primary output) provide minimal benefit while weather and schedules drive most gains; (2) whether the sampler comparison conflates variable-selection quality with the universal benefit of weather; and (3) the absence of error bars or significance tests. We address each below. We agree with all three points and will revise accordingly.","responses":[{"response":"The referee is correct that the current ablation does not isolate the agentic sampler's contribution from the universal benefit of weather and schedules. This is a genuine gap in the experimental design, and we will address it by running the requested ablation: weather forecasts and operational schedules will be added on top of alternative past-known sensor selection strategies (random, k-hop, same-ontology) to test whether the agentic sampler's past-known sensor selections provide additional benefit beyond what weather and schedules alone deliver. We will also add a condition where weather and schedules are used with no past-known sensors at all, establishing a clean baseline for the availability-aware formulation independent of any sampler. If the results show that the sampler's past-known selections do not add significant value beyond weather and schedules, we will explicitly reframe the central claim. Specifically, we will reposition the paper's contribution as: (i) the availability-aware past/future-known formulation, which is the mechanism that enables weather and schedule injection in a deployment-faithful manner, and (ii) the agentic topology sampler as a principled variable-selection method that is especially valuable for physically coupled HVAC variables (as supported by the per-ontology analysis in Table 3), even if its aggregate-level contribution is smaller than that of future-known exogenous variables. We agree that the current framing overstates the sampler's role in aggregate performance and will revise the abstract, introduction, and conclusion to accurately reflect the relative contributions. We will not claim that the sampler is the primary driver of aggregate gains unless the new ablation supports that claim.","revision_made":"yes","referee_comment":"§4.4, Figure 3: The ablation reveals that +PastObs—the primary output of the agentic topology sampler—provides minimal or negative benefit (+4.7% nMAE on BTS-B at H=24, +10.0% at H=48), while gains come primarily from +WeatherFcst and +OpSched. Since weather and schedules are building-level global variables that do not require target-specific topology reasoning to select, this raises a question about whether the agentic topology sampler—the paper's central novel contribution—is actually responsible for the performance gains. The paper should either (a) run an ablation where weather and schedules are added without the agentic sampler (e.g., using all available past-known sensors selected by random or k-hop) to isolate the sampler's contribution, or (b) explicitly reframe the central claim to acknowledge that the availability-aware formulation and weather injection are the primary drivers."},{"response":"The referee is right that the sampler comparison in Figure 2 does not decompose the advantage by covariate group, and this conflation undermines the claim that topology-aware sampling is more reliable than simpler alternatives. We will run the requested decomposition. Specifically, we will report sampler comparisons (agentic vs. random vs. same-ontology vs. k-hop) separately for: (a) past-known sensors only (no weather, no schedules), (b) future-known variables only (weather + schedules, no past-known sensors), and (c) the full combination. This will reveal whether the agentic sampler's advantage in Figure 2 comes from better past-known sensor selection or from consistently including weather and schedule variables that simpler alternatives might miss. We note that in the current experimental design, all samplers receive oracle weather, so the future-known component should be identical across samplers—meaning any difference in the current Figure 2 should already reflect past-known selection quality. However, the referee's concern is valid because the current presentation does not make this explicit, and we have not verified that all samplers include the same future-known variables. We will add the decomposition and, if it reveals that the sampler's advantage is primarily in consistently including weather/schedules rather than in topology-aware sensor selection, we will revise the claim accordingly. The claim that 'topology-aware sampling is more reliable than random, ontology-only, or fixed-hop selection' will be qualified to specify the covariate group where the advantage holds.","revision_made":"yes","referee_comment":"§4.3, Figure 2: The sampler comparison uses oracle weather for all methods but does not decompose whether the agentic sampler's advantage over random/ontology/k-hop comes from selecting better past-known sensors or better future-known variables. Given that §4.4 shows past-known sensors provide minimal benefit, the sampler comparison may conflate the sampler's variable-selection quality with the universal benefit of weather. A decomposition of the sampler comparison by covariate group (past-known vs. future-known) would clarify whether the agentic sampler's advantage is actually in topology-aware sensor selection or simply in consistently including weather/schedule variables that simpler alternatives might miss. This is load-bearing for the claim that topology-aware sampling is more reliable than simpler alternatives."},{"response":"The referee is correct. We will add bootstrap confidence intervals and paired tests across per-sensor results for the key comparisons in Table 2. Specifically, we will report 95% bootstrap confidence intervals for nMAE and nMSE for TopoBrick and the strongest baselines at each building/horizon setting, and we will add paired comparisons (e.g., paired t-tests or Wilcoxon signed-rank tests across per-sensor errors) for the closest-margin comparisons the referee identifies, including TopoBrick vs. PatchTST on BTS-C at H=24 and H=72. Where differences are not statistically significant, we will state this explicitly and avoid claiming superiority. We will also add a note in the results discussion acknowledging which comparisons are within noise. This is a straightforward addition that does not change the experimental pipeline, only the reporting.","revision_made":"yes","referee_comment":"Table 2: No error bars, confidence intervals, or statistical significance tests are reported for any result. Several key comparisons involve small margins (e.g., TopoBrick nMAE 0.319 vs. PatchTST 0.319 on BTS-C at H=24; TopoBrick nMAE 0.476 vs. PatchTST 0.473 on BTS-C at H=72). Without variance estimates, it is unclear whether these differences are meaningful. At minimum, bootstrap confidence intervals or paired tests across the per-sensor results should be reported for the key comparisons."}],"tokens_in":20504,"tokens_out":1440,"duration_ms":423255,"standing_objections":[]},"desk_editor":{"model":"glm-5.2","letter":"The paper you should know about: TopoBrick uses an LLM-based agentic sampler over building knowledge graphs (Brick schema) to select target-specific exogenous variables for zero-shot forecasting, organized by deployment-time availability (past-known sensors vs. future-known weather/schedules). The results on three real buildings across two continents are solid, with TopoBrick beating zero-shot foundation models (Chronos-2, Moirai, TimesFM) and matching or beating fully supervised baselines on two of three buildings. Code is public, which is a real plus. The deployment-faithful past/future-known split is a genuinely useful formulation that other building forecasting work should adopt. The per-ontology breakdown (Table 3) is informative and honest about where the method helps (HVAC-coupled, weather-driven sensors) and where it doesn't (electrical demand, airflow). The sampler comparison in §4.3 showing the agentic approach beating random, ontology-only, and k-hop alternatives is convincing. Now the soft spot, and it's a real one. The stress-test concern lands. The ablation in §4.4 shows that +PastObs—the topology-selected sensor variables that are the agentic sampler's primary output—provide minimal or negative benefit (+4.7% nMAE on BTS-B at H=24, +10.0% at H=48). The gains come from +WeatherFcst and +OpSched, which are building-level global variables that don't require target-specific topology reasoning to select. You just include all weather channels and all schedules. This creates genuine tension with the paper's central framing of topology-aware sampling as the key innovation. The §4.3 sampler comparison partially addresses this—the agentic sampler does beat alternatives—but that experiment uses oracle weather for all methods and doesn't decompose whether the sampler's edge comes from better sensor selection or better future-known variable selection. The paper would be much stronger with one additional experiment: the same ablation as §4.4 but comparing agentic vs. random vs. k-hop specifically on the +PastObs component, to show the sampler's sensor selections matter even if the aggregate PastObs contribution is small. Minor issues: no error bars or significance tests, and the LLM (gpt-oss-20b) is unfamiliar with no variance reported across agent runs. These don't sink the paper but should be addressed. The central contribution still holds in a qualified form: the availability-aware formulation plus topology-grounded variable selection is a practical, training-free pipeline that works. But the paper overclaims when it attributes gains primarily to topology-aware sampling rather than to the future-known covariate formulation. This deserves a serious referee. The core idea is good, the experiments are mostly fair, and the code is available. A referee should push the authors to either show that the agentic sampler's sensor selections specifically contribute (with the targeted ablation above) or reframe the contribution to center on the availability-aware formulation rather than the sampler itself.","headline":"Training-free zero-shot building IoT forecasting using LLM agents over Brick knowledge graphs to select exogenous variables; competitive with supervised baselines on three real buildings.","tokens_in":21628,"tokens_out":689,"would_cite":true,"duration_ms":291808,"reading_group":"no","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"glm-5.2","headline":"LLM picks building sensors by physical topology, beats zero-shot baselines","keywords":["zero-shot forecasting","building knowledge graph","exogenous variable selection","agentic reasoning","Brick Schema","building IoT","time series foundation models","topology-aware forecasting"],"falsifier":"If a simpler pipeline using the same future-known variables (weather forecasts, operational schedules, calendar features) but with random or fixed-hop past-known variable selection achieves forecasting performance comparable to TopoBrick, then the agentic topology sampler is not the load-bearing mechanism.","tokens_in":20593,"feed_emoji":"","tokens_out":1181,"duration_ms":249218,"temperature":0.7,"pith_summary":"TopoBrick uses an LLM-based agentic topology sampler over building knowledge graphs to select target-specific exogenous variables for zero-shot building IoT forecasting. The framework constructs a compact building skeleton from the raw knowledge graph, reasons over target-centric topology context to identify physically and operationally relevant exogenous variables, verifies selections against graph evidence, and organizes selected variables by deployment-time availability (past-known sensor states vs. future-known calendar, schedule, and meteorological variables). The central claim is that topology-aware exogenous-variable selection provides a stronger inductive bias than either channel-independent zero-shot forecasting or per-building supervised training, as evidenced by best-or-second-best normalized MAE and MSE across three real-world buildings on two continents. The paper demonstrates that a frozen time-series foundation model, when supplied with topology-routed exogenous variables organized by availability, can match or exceed fully trained building-specific models without any building-specific training. The agentic sampler outperforms random, ontology-only, and fixed-hop selection strategies, particularly for HVAC and weather-driven sensing variables. However, ablation results reveal that future-known meteorological forecasts and operational schedules drive most of the performance gains, while topology-selected past-known sensor observations provide minimal benefit and sometimes degrade performance.","feed_headline":"","feed_subtitle":"","key_machinery":"The central mechanism is the agentic topology sampler, which operates on a compact building skeleton distilled from the raw knowledge graph. Given a target Point node, the sampler receives a target-centric topology context comprising three components: the target anchor (equipment, room, zone, or system to which the target is attached), the local topology (upstream equipment, downstream served entities, sibling equipment, spatial containers), and global context (building-level drivers such as weather points and operational schedules). The agent reasons over this context following three principles—physical relevance (identifying plausible driver categories for the target's physical quantity), ","core_discovery":"The paper's central discovery is that a building knowledge graph can serve as an effective routing layer for exogenous-variable selection in zero-shot forecasting, enabling a frozen foundation model to leverage building-specific physical and operational context without training. The agentic topology sampler—reasoning over equipment flow, spatial containment, and control-loop structure—selects more informative exogenous variables than graph-distance heuristics or ontology-class matching. The availability-aware formulation, which separates past-known sensor states (masked over the prediction horizon) from future-known calendar, schedule, and meteorological variables, proves critical: future-已知","pith_inferences":["The ablation results raise the possibility that the agentic topology sampler—the paper's central contribution—may not be the primary driver of performance improvements. If future-known weather forecasts and operational schedules account for most gains, a simpler pipeline that injects these variables without topology-based selection might achieve comparable results, which would narrow the contribut","A direct test of this concern would compare TopoBrick against a baseline that uses the same future-known variables (weather, schedules, calendar) but selects past-known exogenous variables randomly or via fixed-hop sampling rather than agentic topology reasoning. If performance differences are negligible, the agentic sampler's value is called into question.","The ontology-level results suggest that the framework's effectiveness is bounded by whether the true physical drivers of a target sensor are represented in the knowledge graph. For variables dominated by unobserved occupancy, stochastic equipment usage, or closed-loop control policies, topology-aware selection may have limited leverage regardless of sampling quality.","The reliance on an LLM (gpt-oss-20b) for agentic reasoning introduces a dependency on model quality and prompt design. The paper does not report sensitivity to LLM choice or reasoning failures, leaving open whether the framework's robustness holds across different language model backbones."],"forward_implications":["Building knowledge graphs could serve as routing layers for other building analytics tasks beyond forecasting, including virtual sensing, anomaly detection, fault localization, and sensor outage recovery.","The availability-aware formulation (past-known vs. future-known) could be adopted by existing time-series foundation models to improve deployment-faithful forecasting in cyber-physical systems more broadly.","The finding that future-known weather and schedules drive most gains suggests that improving weather forecast quality and schedule integration may yield more returns than refining topology-based sensor selection.","The framework's training-free nature makes it deployable across heterogeneous building portfolios without per-building data collection or model tuning, potentially lowering the barrier to scalable building analytics."],"fun_headline_variants":["TopoBrick: Agentic topology sampling for zero-shot building IoT forecasting","Zero-shot building IoT forecasting via agentic topology sampling","Building knowledge graphs route variables for zero-shot IoT forecasting","Training-free zero-shot forecasting using agentic topology sampling","TopoBrick leverages building graphs for zero-shot IoT forecasting"],"cache_read_input_tokens":0,"weakest_assumption_plain":"The paper assumes that LLM-based agentic reasoning over building topology context can reliably identify physically relevant exogenous variables that improve forecasts, but this assumption is validated only indirectly through downstream forecasting metrics. The ablation reveals that topology-selected past-known sensor observations provide minimal benefit or even degrade performance, while future-known weather forecasts drive most gains—raising the question of whether the agent","fun_headline_variants_meta":{"raw":{"variants":["TopoBrick: Agentic topology sampling for zero-shot building IoT forecasting","Zero-shot building IoT forecasting via agentic topology sampling","Building knowledge graphs route variables for zero-shot IoT forecasting","Training-free zero-shot forecasting using agentic topology sampling","TopoBrick leverages building graphs for zero-shot IoT forecasting","Exogenous variable routing via building graphs for zero-shot forecasting"]},"model":"glm-5.2","effort":"high","cost_usd":0.0,"raw_usage":{"total_tokens":1054,"prompt_tokens":427,"completion_tokens":627,"prompt_tokens_details":null},"tokens_in":427,"tokens_out":627,"duration_ms":24601,"temperature":1.0,"reasoning_tokens":604,"cache_read_input_tokens":0,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-08T08:42:14.693880+00:00","model_set":{"reader":"glm-5.2"},"falsifier":"If a simpler pipeline using the same future-known variables (weather forecasts, operational schedules, calendar features) but with random or fixed-hop past-known variable selection achieves forecasting performance comparable to TopoBrick, then the agentic topology sampler is not the load-bearing mechanism.","supporting_citations":[],"review_version":1}