Pith. sign in

REVIEW 5 major objections 6 minor 34 references

Parking, Perception, and Retail: Street-Level Determinants of Community Vitality in Harbin

T0 review · 5 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read The paper claims that on-street parking helps or hurts neighborhood commerce depending on street width, with excessive parking on narrow streets lowering both satisfaction and shop prices.

desk verdict A promising pilot that currently hides its evidence: no regression tables, a hand-tuned outcome index, and contradictory model attributions make the central claims unverifiable as written. read the letter →

arxiv 2506.05080 v1 pith:YXSBLUVZ submitted 2025-06-05 cs.CL cs.CV

HaoTian Lan This is my paper
classification cs.CLcs.CV
keywords commercialvitalityon-streetparkingstreetviewimagerymultimodallargelanguagemodelwalkabilitygatedcommunitiesHarbinstreetscapequality
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper is trying to establish that the relationship between parked cars and neighborhood commercial vitality is not monotonic: it is conditional on street geometry. Using Harbin's gated-community perimeter streets, it finds that a moderate level of vehicle presence can coincide with stronger retail activity, but on narrow streets additional on-street parking is associated with lower resident satisfaction and lower shop-level prices, while on wide streets prices rise with parking up to a threshold and then fall. It also argues that perceived greenery and cleanliness are the strongest correlates of pedestrian satisfaction yet only weak correlates of shop pricing, whereas street and sidewalk width correlate more strongly with pricing. If the paper is right, parking policy should be tailored to street type rather than governed by the simple assumption that more parking always helps retail. The paper also demonstrates that street-view imagery plus a multimodal language model can generate usable streetscape indicators at community scale.

What carries the argument

The analytical engine is a regression model in which the average number of parked vehicles per street-view image is interacted with street width, so that the effect of parking is allowed to differ between narrow and wide streets. The outcome side is the Community Commercial Vitality Index (CCVI), a weighted sum of business counts, ratings, review volumes, and anchor amenities built from Meituan and Dianping data, plus separate measures of average store price and perceived satisfaction. The input side comes from roughly 10–15 Baidu Street View panoramas per community, parsed by a multimodal large language model (VisualGLM-6B) with GPT-4-based prompting and human-annotated calibration, yielding vehicle counts, parking presence, greenery levels, width categories, and cleanliness scores. This machinery lets the paper turn free-text image descriptions into structured variables and test whether the parking effect is moderated by street geometry.

What would settle it

A reader could settle the claim by taking a held-out set of Harbin communities, having human auditors count parked vehicles and rate greenery, cleanliness, and width from the same Baidu panoramas, and re-running the regressions with a PCA-weighted CCVI built from raw business data. If the AI-derived counts disagree with human counts in a way that correlates with vitality, or if the vehicle–width interaction flips sign once the index is reweighted, the claimed moderation would be a measurement artifact.

Watch

Extended reading notes

Core claim

The central discovery is a conditional, width-moderated relationship between on-street vehicle presence and commercial outcomes. In communities with narrow streets, an increase in parked vehicles is associated with lower satisfaction and lower average store prices, likely through congestion and reduced walkability. On wider streets, average store prices initially rise as parking increases, peak around 50–60 parking spaces, and then decline, while satisfaction declines slightly with vehicle density. Greenery and cleanliness predict perceived satisfaction most strongly but have weak associations with pricing, and sidewalk and carriageway width are the stronger correlates of store pricing. The paper reads this as evidence that micro-scale street design can activate or suppress commercial potential even after macro factors like distance to the central business district and population density are controlled.

Load-bearing premise

The load-bearing premise is that the Community Commercial Vitality Index and the AI-parsed street variables actually measure what they stand for: the CCVI weights were set by empirical judgment and iteratively adjusted to match qualitative impressions, and the vehicle, greenery, cleanliness, and width variables come from unvalidated text parsing of a small number of images per community.

Editorial extensions

If this is right

  • On narrow commercial streets, reducing curbside parking or moving vehicles to off-street lots should protect both pedestrian satisfaction and shop-level pricing.
  • On wide commercial corridors, parking provision can support higher shop prices only up to a saturation point beyond which further parking is associated with decline.
  • Greenery and cleanliness investments are likely to raise residents' satisfaction even when they do not move commercial price positioning.
  • Street width should be recorded as a standard moderator in any street-level study of parking and retail outcomes.
  • AI-based street-view audits can be scaled to many communities, making context-sensitive parking diagnostics feasible for planning practice.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the width moderation generalizes, other Chinese cities with gated-community layouts should show the same sign pattern, so a cross-city replication would be a direct test of the paper's mechanism.
  • The reported peak in the wide-street parking–price curve is a quantitative prediction: an optimal on-street parking density near 50–60 spaces per sampled frontage, which could be tested against independent transaction data.
  • Because the paper's outcome variables are cumulative review counts and listed prices, a stronger test would use time-stamped foot traffic or sales records to see whether the same threshold appears in actual spending behavior.
  • The AI-parsing step is a potential source of artifact, so a validation experiment comparing GPT-4/VisualGLM reads against manual counts on held-out images would show whether measurement error is correlated with vitality.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

5 major / 6 minor

Summary. This paper investigates how street-level features - parked vehicle density, greenery, cleanliness, and street width - relate to commercial vitality in Harbin's gated communities. Using Baidu Street View imagery, the authors extract streetscape attributes with multimodal large language models (GPT-4/VisualGLM-6B) and construct a Community Commercial Vitality Index (CCVI) from Meituan and Dianping data. The central claim is a conditional, non-linear relationship: moderate vehicle presence supports commerce, but excessive on-street parking - especially on narrow streets - reduces satisfaction and shop-level pricing, while wider streets exhibit an inverted-U pricing response to parking. The paper reports descriptive statistics, bivariate correlations, and narrative regression results, and concludes with policy recommendations for parking management and street design.

Significance. The paper addresses an important gap in the parking-retail literature by examining community-scale streets in a Chinese context, an understudied setting. The use of multimodal LLMs for streetscape feature extraction is a promising methodological direction, and the conditional framing (street width as a moderator) is a useful hypothesis. However, the empirical evidence as presented is insufficient to support the claims: no regression tables are provided, the outcome variables 'satisfaction' and 'average store price' are not defined, and the hand-tuned CCVI introduces circularity. The paper's own conclusion acknowledges that robustness and causal channels are unconfirmed, yet the abstract states the findings as established. As such, the contribution is currently more of a research proposal than a verified empirical study.

major comments (5)
  1. [Section 4.3, Eq. (2)] The entire conditional claim - negative parking effects on narrow streets and an inverted-U on wide streets - rests on regression results that are never reported. Provide the full estimation output for Eq. (2) and for the interaction model behind Figure 5, including coefficients, standard errors, t-statistics, p-values, confidence intervals, sample size, and adjusted R-squared. Without these, the reader cannot verify the strength or even the existence of the reported effects.
  2. [Section 3.2.2, Eq. (1)] The CCVI weights are 'determined based on empirical judgment and iteratively adjusted to align with qualitative assessments' of which communities are vibrant. This means the dependent variable is co-constructed with the researchers' prior ranking, so correlations between streetscape features and CCVI may be circular. Demonstrate that the results are robust to alternative weighting schemes, or use a data-driven method such as PCA, and report the resulting weights.
  3. [Section 3.2.2 vs. Sections 4.2-4.4] 'Satisfaction' and 'average store price' are used as key outcomes in the abstract and in Sections 4.2-4.4, but Section 3.2.2 lists only total businesses, review counts, average ratings, sales volume, and anchor presence as the CCVI components. Define how satisfaction scores and average store prices were operationalized, including their data sources, units, and aggregation to community level.
  4. [Sections 3.2.3, 5.5, 6] The feature extraction model is described inconsistently: Section 3.2.3 says GPT-4 was queried with a fine-tuned prompt, while Section 5.5 and Section 6 attribute the analysis to multimodal models 'such as VisualGLM', and the abstract names VisualGLM-6B. Clarify which model(s) produced the vehicle counts, greenery categories, width, and cleanliness scores, and how they were combined; this is essential for reproducibility.
  5. [Section 6, Conclusion] The conclusion states that 'our current model cannot yet confirm their robustness or isolate causal channels' and calls the results preliminary. This directly contradicts the abstract's assertion that 'excessive on-street parking - especially in narrow streets - erodes walkability and reduces both satisfaction and shop-level pricing.' Align the abstract with the actual strength of the evidence, or provide the missing robustness checks.
minor comments (6)
  1. [Section 3.2.2, Eq. (1)] Equation (1) is incomplete: the weights are shown as placeholder symbols ('?1', '. . .') and the index components are not exhaustively specified. Complete the equation and state the range and direction of each component.
  2. [References] The reference list contains several duplicates (e.g., Hymel 2014 as refs 11 and 25; Merten & Kuhnimhof 2023 as refs 10 and 26; Shoup as refs 12 and 27; Appleyard as refs 13 and 28; Li et al. 2024 as refs 15 and 24; Zhang & Hu 2022 as refs 22 and 37). Consolidate duplicate citations to improve clarity.
  3. [Figure 5] Figure 5 is presented without the underlying data points, confidence bands, or a statement of the fitted functional form. Add this information or refer to a table with the regression estimates.
  4. [Section 4.1] Section 4.1 introduces 'perceived satisfaction' and 'average store price' before these constructs are defined in the methodology; move their definitions to Section 3.2.2 or add a dedicated variables subsection.
  5. [Table 1] Table 1 reports average vehicle counts as non-integer values; add a note explaining that these are per-image averages across the 10-15 images per community.
  6. [Abstract and Section 2.3] The paper alternates between 'multimodal large language models (MLLMs)' and 'multimodal large models (MLLMs)' in the abstract and Section 2.3; standardize the terminology.

Circularity Check

1 steps flagged · score 4.0 of 10

One circular validation loop: CCVI is calibrated to qualitative judgments and then 'externally' validated against the same kind of qualitative knowledge; the central streetscape–price regressions are not circular by construction.

  1. fitted input called prediction [Section 3.2.2 (Eq. 1 construction) and Section 3.3 (validation)]
    "Weights ?? were determined based on empirical judgment and iteratively adjusted to align with qualitative assessments (e.g., communities known locally for vibrant commerce consistently achieved higher CCVI scores). ... Externally, we cross-referenced model-predicted vitality with qualitative local knowledge, such as media reports or known hotspots of street activity in Harbin."

    The CCVI outcome is not an independent measurement: its weights are tuned until communities the authors already judge as vibrant receive high scores. The paper then fits a model to predict this CCVI and validates the predictions by comparing them with qualitative local knowledge, which is the same kind of judgment used to set the weights. The external check therefore reduces to checking the model against its own construction target; it provides no independent confirmation. This does not by construction force the streetscape coefficients in Eq. (2), but it invalidates the claimed external robustness of the predicted-vitality check.

full rationale

The central empirical claims about parked vehicles, street width, satisfaction, and shop pricing are not circular in the strict sense: the streetscape features come from image analysis while the outcome data come from Meituan/Dianping, and the paper does not define the street features in terms of the outcomes. The main circularity is confined to the validation loop: CCVI weights are set by qualitative judgment, and the 'external' validation of model-predicted vitality uses the same kind of qualitative knowledge, so that step is self-referential rather than independent. Other weaknesses—such as unoperationalized 'satisfaction' and 'average store price' and the absence of reported coefficient tables—are reproducibility or completeness issues, not reductions by construction. A self-citation chain is not present. Given that the core regression result still depends on real data rather than on the weights alone, a moderate score is appropriate.

Assumptions & free parameters 3 free parameters · 3 assumptions · 1 invented entities

The central claim rests on the validity of a hand-fitted outcome index and unvalidated AI-extracted street features. The CCVI weights are chosen to reproduce the authors' qualitative ranking of vibrant communities, and the image variables come from text parsing with no released prompts, labels, or error analysis. These are load-bearing measurement assumptions, not minor details.

free parameters (3)
  • CCVI component weights (w1, w2, w3, ...) = not reported
    Section 3.2.2, Equation (1): weights were 'determined based on empirical judgment and iteratively adjusted to align with qualitative assessments.'
  • Qualitative-to-numeric mapping rules for GPT-4 text parsing = e.g., 'a few trees' -> 2 on 0-3 scale; 'about four vehicles' -> 4; 'clean with scattered litter' -> 4/5
    Section 3.2.3: hand-coded conversion rules translate AI text into numbers with no validation set or inter-coder agreement reported.
  • Street width category thresholds = narrow <10 m; medium 10-20 m; wide >20 m
    Section 4.1: imposed bins for analysis and moderation; no sensitivity testing reported.
assumptions (3)
  • domain assumption Meituan and Dianping review counts and ratings are valid proxies for commercial vitality.
    Section 3.2.2 uses these platform metrics to build CCVI without external validation against sales or footfall.
  • domain assumption 10-15 summer street view images per community represent the usual street conditions of that community.
    Section 3.2.1 samples images at 50-100 m intervals; Section 5.4 concedes images are time-bound snapshots.
  • domain assumption GPT-4/VisualGLM perceptual outputs approximate human perception of cleanliness, greenery, and vehicle presence.
    Section 3.2.3 reports preliminary consistency but gives no validation metrics; Section 5.5 calls the approach experimental.
invented entities (1)
  • Community Commercial Vitality Index (CCVI)
    purpose: Composite 0-100 score intended to measure overall commercial scale and activity of each community.
    Section 3.2.2: weights are hand-tuned and iteratively adjusted to match qualitative impressions; no external dataset or independent benchmark is used to validate the index.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Parking, Perception, and Retail: Street-Level Determinants of Community Vitality in Harbin." pith.science (2026). https://pith.science/paper/YXSBLUVZ

@misc{pith2026250605080,
  author       = {Pith},
  title        = {Pith review of: Parking, Perception, and Retail: Street-Level Determinants of Community Vitality in Harbin},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YXSBLUVZ}},
  note         = {Machine review of arXiv:2506.05080}
}
read the original abstract

The commercial vitality of community-scale streets in Chinese cities is shaped by complex interactions between vehicular accessibility, environmental quality, and pedestrian perception. This study proposes an interpretable, image-based framework to examine how street-level features -- including parked vehicle density, greenery, cleanliness, and street width -- impact retail performance and user satisfaction in Harbin, China. Leveraging street view imagery and a multimodal large language model (VisualGLM-6B), we construct a Community Commercial Vitality Index (CCVI) from Meituan and Dianping data and analyze its relationship with spatial attributes extracted via GPT-4-based perception modeling. Our findings reveal that while moderate vehicle presence may enhance commercial access, excessive on-street parking -- especially in narrow streets -- erodes walkability and reduces both satisfaction and shop-level pricing. In contrast, streets with higher perceived greenery and cleanliness show significantly greater satisfaction scores but only weak associations with pricing. Street width moderates the effects of vehicle presence, underscoring the importance of spatial configuration. These results demonstrate the value of integrating AI-assisted perception with urban morphological analysis to capture non-linear and context-sensitive drivers of commercial success. This study advances both theoretical and methodological frontiers by highlighting the conditional role of vehicle activity in neighborhood commerce and demonstrating the feasibility of multimodal AI for perceptual urban diagnostics. The implications extend to urban design, parking management, and scalable planning tools for community revitalization.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

34 extracted references · 34 canonical work pages

  1. [1]

    Built environment, urban vitality and social cohesion: Do vibrant neighborhoods foster strong communities?[J]

    Mouratidis K, Poortinga W. Built environment, urban vitality and social cohesion: Do vibrant neighborhoods foster strong communities?[J]. Landscape and Urban Planning, 2020, 204: 103951

    work page 2020

  2. [2]

    Walkability and its relationships with health, sustainability, and livability: elements of physical environment and evaluation frameworks[J]

    Baobeid A, Koç M, Al-Ghamdi S G. Walkability and its relationships with health, sustainability, and livability: elements of physical environment and evaluation frameworks[J]. Frontiers in Built Environment, 2021, 7: 721218

    work page 2021

  3. [3]

    Study on management strategy of the on-street parking in Urban Residential Area-Taking Harbin as an Example[C]//IOP Conference Series: Earth and Environmental Science

    Xiaoyan Z, Xiqiao Z, Xu S. Study on management strategy of the on-street parking in Urban Residential Area-Taking Harbin as an Example[C]//IOP Conference Series: Earth and Environmental Science. IOP Publishing, 2021, 687(1): 012190

    work page 2021

  4. [4]

    Comparing measures of urban land use mix[J]

    Song Y, Merlin L, Rodriguez D. Comparing measures of urban land use mix[J]. Computers, environment and urban systems, 2013, 42: 1-13

    work page 2013

  5. [5]

    Livable streets: protected neighborhoods?[J]

    Appleyard D. Livable streets: protected neighborhoods?[J]. The ANNALS of the American Academy of Political and Social Science, 1980, 451(1): 106-117

    work page 1980

  6. [6]

    Urban Visual Intelligence: Studying Cities with AI and Street-level Imagery

    Zhang F, Miranda A S, Duarte F, et al. Urban visual intelligence: Studying cities with AI and street-level imagery[J]. arXiv preprint arXiv:2301.00580, 2023

    work page Pith review arXiv 2023

  7. [7]

    The rise and fall of land use mix: review and prospects[J]

    Zhuo Y, Jing X, Wang X, et al. The rise and fall of land use mix: review and prospects[J]. Land, 2022, 11(12): 2198

    work page 2022

  8. [8]

    The Scale Effect of Street View Images and Urban Vitality Is Consistent with a Gaussian Function Distribution[J]

    Yin J, Chen R, Zhang R, et al. The Scale Effect of Street View Images and Urban Vitality Is Consistent with a Gaussian Function Distribution[J]. Land, 2025, 14(2): 415

    work page 2025

Show all 34 references
  1. [9]

    Evaluating the impact of land-use density and mix on spatiotemporal urban activity patterns: An exploratory study using mobile phone data[J]

    Jacobs-Crisioni C, Rietveld P, Koomen E, et al. Evaluating the impact of land-use density and mix on spatiotemporal urban activity patterns: An exploratory study using mobile phone data[J]. Environment and Planning A, 2014, 46(11): 2769-2785

    2014

  2. [16]

    Parking space management via dynamic performance-based pricing[J]

    Mackowski D, Bai Y, Ouyang Y. Parking space management via dynamic performance-based pricing[J]. Transportation Research Procedia, 2015, 7: 170-191

    2015

  3. [17]

    Evaluating the effects of cashing out employer-paid parking: Eight case studies[J]

    Shoup D C. Evaluating the effects of cashing out employer-paid parking: Eight case studies[J]. Transport Policy, 1997, 4(4): 201-216

    1997

  4. [18]

    Parking management best practices: Victoria[J]

    Litman T. Parking management best practices: Victoria[J]. 2006

    2006

  5. [19]

    Street vitality: what predicts pedestrian flows and stationary activities on predominantly residential Chinese streets, at the mesoscale?[J]

    Istrate A L. Street vitality: what predicts pedestrian flows and stationary activities on predominantly residential Chinese streets, at the mesoscale?[J]. Journal of Planning Education and Research, 2025, 45(1): 66-80

    2025

  6. [20]

    How do Retail Stores Affect Pedestrian Walking Speed: An Empirical Observation[C]//International Conference on Traffic and Granular Flow

    Li D. How do Retail Stores Affect Pedestrian Walking Speed: An Empirical Observation[C]//International Conference on Traffic and Granular Flow. Singapore: Springer Nature Singapore, 2022: 215-222

    2022

  7. [21]

    The relationship between the quality of active frontages and public perceptions of public spaces[J]

    Heffernan E, Heffernan T, Pan W. The relationship between the quality of active frontages and public perceptions of public spaces[J]. Urban Design International, 2014, 19: 92-102

    2014

  8. [22]

    Spatial relationship between green view index and normalized differential vegetation index within the Sixth Ring Road of Beijing[J]

    Li T, Zheng X, Wu J, et al. Spatial relationship between green view index and normalized differential vegetation index within the Sixth Ring Road of Beijing[J]. Urban Forestry & Urban Greening, 2021, 62: 127153

    2021

  9. [23]

    Spatial pattern and heterogeneity of green view index in mountainous cities: a case study of Yuzhong district, Chongqing, China[J]

    Gou A, Wang X, Wang J, et al. Spatial pattern and heterogeneity of green view index in mountainous cities: a case study of Yuzhong district, Chongqing, China[J]. Scientific Reports, 2025, 15(1): 12576

    2025

  10. [24]

    Crowdsourced reviews reveal substantial disparities in public perceptions of parking[J]

    Li L, Hu S, Dinh L, et al. Crowdsourced reviews reveal substantial disparities in public perceptions of parking[J]. arXiv preprint arXiv:2407.05104, 2024

    2024 arXiv

  11. [25]

    Do parking fees affect retail sales? Evidence from Starbucks[J]

    Hymel K. Do parking fees affect retail sales? Evidence from Starbucks[J]. Economics of Transportation, 2014, 3(3): 221-233

    2014

  12. [26]

    Impacts of parking and accessibility on retail-oriented city centres[J]

    Merten L, Kuhnimhof T. Impacts of parking and accessibility on retail-oriented city centres[J]. Journal of Transport Geography, 2023, 113: 103733

    2023

  13. [27]

    High cost of free parking[M]

    Shoup D. High cost of free parking[M]. Routledge, 2021

    2021

  14. [28]

    Livable Streets University of California Press[J]

    Appleyard D. Livable Streets University of California Press[J]. Los Angeles & London, 1981

    1981

  15. [29]

    Paid parking in cities: regulating parking demand and consequences for the retail sector[J]

    Rietveld P, Koetse M. Paid parking in cities: regulating parking demand and consequences for the retail sector[J]. Studies in Regional Science, 2008, 38(1): 215-225

    2008

  16. [30]

    The sensitivity of on-street parking demand in response to price changes: A case study in Seattle, WA[J]

    Ottosson D B, Chen C, Wang T, et al. The sensitivity of on-street parking demand in response to price changes: A case study in Seattle, WA[J]. Transport Policy, 2013, 25: 222-232

    2013

  17. [31]

    Parking management best practices[M]

    Litman T. Parking management best practices[M]. Routledge, 2020

    2020

  18. [32]

    Street view imagery in urban analytics and GIS: A review[J]

    Biljecki F, Ito K. Street view imagery in urban analytics and GIS: A review[J]. Landscape and Urban Planning, 2021, 215: 104217

    2021

  19. [34]

    Decoding urban landscapes: Google street view and measurement sensitivity[J]

    Kim J H, Lee S, Hipp J R, et al. Decoding urban landscapes: Google street view and measurement sensitivity[J]. Computers, Environment and Urban Systems, 2021, 88: 101626

    2021

  20. [35]

    Assessing urban greenery by harvesting street view data: A review[J]

    Lu Y, Ferranti E J S, Chapman L, et al. Assessing urban greenery by harvesting street view data: A review[J]. Urban Forestry & Urban Greening, 2023, 83: 127917

    2023

  21. [38]

    Street view imagery: Methods and applications based on artificial intelligence[J]

    Zhang F, Yu L I U. Street view imagery: Methods and applications based on artificial intelligence[J]. National Remote Sensing Bulletin, 2021, 25(5): 1043-1054

    2021

  22. [39]

    Assessing street-level urban greenery using Google Street View and a modified green view index[J]

    Li X, Zhang C, Li W, et al. Assessing street-level urban greenery using Google Street View and a modified green view index[J]. Urban Forestry & Urban Greening, 2015, 14(3): 675-685

    2015

  23. [40]

    Using deep learning and Google Street View to estimate the demographic makeup of neighborhoods across the United States[J]

    Gebru T, Krause J, Wang Y, et al. Using deep learning and Google Street View to estimate the demographic makeup of neighborhoods across the United States[J]. Proceedings of the National Academy of Sciences, 2017, 114(50): 13108-13113

    2017

  24. [41]

    Street view imagery: AI-based analysis method and application[J]

    Song H. Street view imagery: AI-based analysis method and application[J]. Applied and Computational Engineering, 2024, 40: 54-62

    2024

  25. [42]

    Cogvlm: Visual expert for pretrained language models[J]

    Wang W, Lv Q, Yu W, et al. Cogvlm: Visual expert for pretrained language models[J]. Advances in Neural Information Processing Systems, 2024, 37: 121475-121499

    2024

  26. [43]

    Analyzing green view index and green view index best path using Google street view and deep learning[J]

    Zhang J, Hu A. Analyzing green view index and green view index best path using Google street view and deep learning[J]. Journal of Computational Design and Engineering, 2022, 9(5): 2010-2023

    2022

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.