Pith. sign in

REVIEW 4 major objections 6 minor 56 references

Estimating Online Influence Needs Causal Modeling! Counterfactual Analysis of Social Media Engagement

T0 review · 4 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read The paper claims that online influence can be measured causally by modeling external temporal signals as treatments, and that the resulting causal effect scores rank influential users better than follower counts.

desk verdict Solid engagement-prediction study wearing a causal-inference costume that doesn't fit; worth peer review, but the ATE claims need serious reframing. read the letter →

arxiv 2505.19355 v1 pith:SYBUIBIO submitted 2025-05-25 cs.CL cs.SI

Lin Tian , Marian-Andrei Rizoiu This is my paper · ORCID
classification cs.CLcs.SI
keywords causalinferencesocialmediainfluencecounterfactualanalysisaveragetreatmenteffectmisinformationengagementpredictionMambaG-computation
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

The paper sets out to establish that true online influence can be measured causally: external temporal signals such as Google Trends spikes are treated as time-varying treatments, and a user's influence is the Average Treatment Effect of their posting activity on later engagement. To do this it adapts a joint treatment-outcome framework from healthcare causal inference to discrete-time social media data, fitting deep sequential models that jointly predict treatment intensity and engagement. Reported results on misinformation and disinformation datasets show 15–22% better engagement prediction than existing benchmarks under counterfactual scenarios that vary exposure, timing, and duration. A case study on 492 users reports that the resulting causal-effect scores agree with the expert-based empirical influence better than follower counts do, and that low-baseline misinformation narratives respond super-linearly to external promotion.

What carries the argument

The load-bearing mechanism is the joint treatment-outcome model: a deep sequential architecture with two heads, one predicting the binary event intensity of external signals (a square-transformed function of baseline, past treatments, past outcomes, and Google Trends windows) and one predicting engagement, trained with a combined loss of mean squared error and binary cross-entropy plus a temporal-coherence or attention-consistency regularizer. Causal quantities are obtained by G-computation: after applying the counterfactual manipulation function $\Psi_\theta$ that shifts each signal's timing by $\delta_\theta$ and scales its intensity by $\gamma_\theta$, the Average Treatment Effect is $\Delta_C = \mathbb{E}[Y \mid G_C] - \mathbb{E}[Y \mid G]$. The instantiation that carries the results is causal-Mamba—Mamba, a selective state-space sequence model, augmented with low-rank adapters that condition state transitions on external signal intensities.

What would settle it

Carry out a randomized field experiment in which identical posts receive randomly assigned exposure intensities or timings (for example, a platform boosts some posts with promotion while others are held out), and compare the experimentally measured effect of the exposure change with the ATE estimated from passive Google Trends data; systematic divergence would refute the causal claim. A cheaper check is to rerun the model across a known feed-ranking algorithm change and look for a sudden jump in prediction error, which would indicate an unmodeled hidden pathway.

Watch

Extended reading notes

Core claim

The paper's central discovery claim is that counterfactual prediction differences, computed by transforming observed Google Trends signals through a temporal manipulation function and taking the difference in expected engagement, are valid Average Treatment Effects of external signals on engagement. Under three stated assumptions—consistency, fully-mediated policy effect, and temporal precedence—the joint treatment-outcome model (dubbed causal-Mamba in its best-performing Mamba-plus-adapter instantiation) is presented as identifying these effects from observational data. The evidence offered is threefold: the model outperforms baselines by 15–22% in engagement prediction; ATEs grow super-linearly with exposure, rise with earlier timing, and increase with longer duration; and causal-effect influence scores align with the expert-based gold standard on 492 users (Spearman $\rho = 0.57$, Kendall's $W = 0.70$, CCC $= 0.21$), whereas follower counts align more weakly ($\rho = 0.49$, $W = 0.67$, CCC $= 0.00$). The paper's own conclusion is that causal effect is a tighter approximation of influence than account popularity.

Load-bearing premise

The load-bearing premise is that external signals such as Google Trends changes affect engagement only through the observable pathways the model captures, with no hidden routes such as algorithmic amplification or coordinated network activity; if hidden confounders exist, the estimated Average Treatment Effects are biased and the resulting influence ranking is not truly causal.

Editorial extensions

If this is right

  • If the causal interpretation holds, counterfactual engagement forecasts become usable for policy design: platforms could estimate the engagement reduction from lowering exposure to a misinformation signal before intervening.
  • The ATE-based influence score can be computed at scale from passive observations, removing the bottleneck of the expert-based empirical influence, which currently exists for only 492 users.
  • Moderation and platform governance could prioritize users, pages, and groups by measured causal effect rather than follower counts, which the paper finds capture a fundamentally different and weaker signal of influence.
  • The reported super-linear response of low-baseline narratives to external signals implies that small Google Trends spikes can trigger disproportionately large engagement bursts, so early intervention timing matters more for low-prominence misinformation than for already-popular content.

Reading between the lines

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

  • Editorial inference: The alignment with expert-influence rankings was tested on 492 users discussing anti-climate topics; the paper does not test other topics, so a natural extension is to collect expert rankings for vaccination or election narratives and check whether the ATE ranking still dominates follower counts there.
  • Editorial inference: The super-linear ATE pattern suggests an epidemic-threshold structure in attention-driven engagement; one could extend the framework to predict ex ante which low-baseline narratives will cross the amplification threshold under a forecast Google Trends shock, which the paper does not formulate.
  • Editorial inference: The strongest untested validation would be a randomized exposure experiment—staggered or boosted promotion of identical posts—comparing experimentally measured effects with the model's passive-data ATEs; the paper reports no such comparison.
  • Editorial inference: Because Assumption 2 excludes algorithmic amplification as a hidden pathway, the framework could be stress-tested by locating known platform ranking-algorithm changes and checking whether causal-Mamba's prediction error jumps at those change points; an error jump would signal an unmodeled confounder.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 6 minor

Summary. The paper proposes a joint treatment-outcome framework that couples deep sequential models (Transformers and Mamba) with Google Trends signals to predict social media engagement under counterfactual manipulations of those signals. The authors define a counterfactual scenario as a temporal transformation of the Google Trends input, compute the difference between model predictions under transformed and observed inputs, and label this difference an Average Treatment Effect (ATE) obtained by G-computation. They report predictive improvements over baselines across several counterfactual scenarios, and a case study on 492 X/Twitter users in which their 'causal effect influence' scores correlate more strongly with an expert-based empirical influence measure than do follower counts. The paper concludes that the approach offers a reliable, scalable causal measure of online influence.

Significance. The intended contribution is substantial: a scalable causal influence measure and a counterfactual prediction tool for misinformation intervention would be valuable, and the paper addresses a real gap in social media influence research. Strengths include the use of multiple real-world misinformation/disinformation datasets (SocialSense, DiN), the systematic comparison of eight architectural variants, bootstrap confidence intervals in Table 3, and a direct comparison against a human-judgment-based influence benchmark. However, the causal identification is not established, and the key empirical validation is incompletely documented. The significance of the paper therefore depends on a load-bearing causal claim that the current evidence does not support.

major comments (4)
  1. [Section 3.2 and Table 3] The paper's central claim that the counterfactual prediction gaps in Table 3 are valid Average Treatment Effects requires sequential exchangeability, i.e., no unobserved confounding given the modeled history, the condition that licenses G-computation. Assumption 2 (Fully-Mediated Policy Effect) is the only place this condition is addressed, and it is defended by predictive performance ('superior predictive performance' of OMM and HIP), which cannot rule out hidden pathways. Section 6 then concedes 'unobserved confounding from algorithmic amplification' and 'treatment interference through network effects,' directly contradicting Assumption 2. No placebo-signal test or sensitivity analysis is provided, so the ATEs in Table 3 and the 'causal effect influence' scores in Section 5 are not identified as causal effects.
  2. [Section 3.3 and Table 3] The estimand is stated as ΔC = E[Y|G_C] − E[Y|G], which is a conditional outcome regression under a transformed input signal. The 'G-computation' label in Table 3 is not supported by the described procedure: there is no explicit marginalization over the engagement history H, no forward simulation of the treatment process under the counterfactual policy, and no demonstration that the outcome model is used as an estimator of the g-formula. Either a proper G-computation estimator should be implemented and described, or the quantity should be relabeled as a model-based counterfactual prediction and the causal language should be adjusted accordingly.
  3. [Section 4.3, Tables 1 and 2] Tables 1 and 2 report single-point RMSE and BCE values without error bars, confidence intervals, or significance tests, even though Table 3 reports bootstrap intervals. The abstract's claim of '15–22%' improvement over existing benchmarks is not directly recoverable from the reported numbers (for example, Mamba+Adapter has base RMSE 0.113 versus MBPP and Transformer at 0.193, which is a much larger margin), and no statistical support is given for the ranking of architectures. This weakens the conclusions drawn for RQ1 and RQ2.
  4. [Section 5] The case study on 492 X/Twitter users does not document how the model's inputs are constructed: which Google Trends keywords are used for these users, how each user's posting behavior is encoded as a treatment signal, what observation and prediction windows are used, and how the per-user causal effect score is aggregated across posts. The empirical influence scores come from a separate human-judgment study on anti-climate-change topics, and the matching procedure between that study's users and the model's input data is not described. Without this documentation, the reported correlations (ρ=0.57 vs. 0.49, W=0.70 vs. 0.67, CCC=0.21 vs. 0.00) cannot be reproduced or assessed.
minor comments (6)
  1. [Section 2] The sentence 'These signals, collected over l time points, serve as an exogenous signal, capturing how real-world interest.' is incomplete and should be finished, for example with 'evolves over time.'
  2. [Section 5] The text refers to 'Fig. 4' when describing the decile heatmaps of Spearman correlation, but the heatmaps appear as Figure 3; later, 'Figure 4' is used for the relative percentage changes in engagement. The figure numbering should be made consistent.
  3. [Section 4.3] The phrase 'Mamba+Adaptors, dubbed causal-Mamba' should read 'Mamba+Adapter' to match the naming used elsewhere.
  4. [Tables 1 and 2] The column header 'Scenario 3: Treatment' in Table 2 should be 'Scenario 3: Duration' to match Table 1.
  5. [Appendix A.2] The hyperparameters α=0.5 and β=0.1 appear only in the appendix; β is not defined in the main text's loss function or in Section 3.4, leaving the temporal-coherence and attention-consistency losses underspecified.
  6. [Figure 3] The caption contains the typo 'Effect Effect Scores'; this should be 'Causal Effect Scores'.

Circularity Check

3 steps flagged · score 5.0 of 10

The Table 3 'ATE' is exactly the fitted model's prediction difference E[Y|GC]-E[Y|G]; its causal reading rests on Assumption 2, whose only cited quantitative support is the authors' own OMM/HIP predictive fit and which §6 contradicts.

  1. self citation load bearing [Section 3.2, Assumption 2 (Fully-Mediated Policy Effect)]
    "While alternative pathways might exist–such as algorithmic amplification or coordinated network activity–these appear to be minimal, as evidenced by both OMM's and HIP's superior predictive performance when directly modeling this attention-mediated relationship."

    This is the identifying assumption that turns E[Y|GC]-E[Y|G] into an ATE. Its only quantitative support is the predictive fit of OMM (Calderon et al. 2024) and HIP (Rizoiu et al. 2017), both from the authors' own group; Section 6 then concedes 'unobserved confounding from algorithmic amplification', exactly the hidden pathway the assumption denies. The causal premise is therefore carried by a self-citation chain plus an inference from predictive fit to no-confounding, not by any test of sequential exchangeability.

  2. renaming known result [Section 3.3 (definition of ΔC) and Table 3 (ATE via G-computation)]
    "we estimate the expected engagement outcomes under the transformed signal and calculate the causal effect as the difference between counterfactual and actual outcomes: ΔC = E[Y|GC] − E[Y|G]."

    By construction, ΔC is the trained model's prediction under transformed Google Trends minus its prediction under observed Google Trends. Table 3 relabels this predictive contrast as 'Average Treatment Effect (ATE), computed via G-computation (Robins, 1986)', but no g-formula, no marginalization over time-varying histories, and no exchangeability condition is derived. The causal 'prediction' is thus equivalent, by definition, to the model's input sensitivity, with causal content imported only by the self-cited Assumption 2.

1 more flagged steps
  1. self citation load bearing [Section 3.2, Assumption 1 (Consistency)]
    "They prove that HIP has the Linear Time-Invariant property which ensures that identical external stimuli produce consistent engagement responses regardless of when they occur."

    Consistency is an assumption about the real-world data-generating process, but the cited 'proof' is a mathematical property of the authors' own HIP model (Rizoiu et al. 2017). Importing a model-theoretic invariant as empirical evidence for the consistency of potential outcomes is a category error, and it is load-bearing because Assumption 1 is needed for the counterfactual interpretation of Section 3.3.

full rationale

The paper's architecture comparisons (Tables 1-2 and the appendix) are self-contained empirical results and are not circular. The circularity is concentrated in the causal interpretation. Section 3.3 defines the causal effect as ΔC = E[Y|GC] − E[Y|G], i.e., exactly the trained model's prediction difference under two input configurations; Table 3 then labels this 'Average Treatment Effect, computed via G-computation'. No g-formula or sequential-exchangeability condition is derived or tested, so the 'ATE' is the model's sensitivity by construction, and causal content comes only from Assumptions 1-3. Assumption 2, the critical no-hidden-pathway condition, is justified by 'superior predictive performance' of OMM and HIP, both previous papers by the same group (Calderon et al. 2024; Rizoiu et al. 2017). Predictive fit cannot rule out unobserved confounding, and the paper's own conclusion (§6) lists 'unobserved confounding from algorithmic amplification' and 'treatment interference through network effects' as limitations, contradicting the assumption's denial of hidden pathways. Assumption 1 is similarly 'supported' by the Linear Time-Invariant property proved for the authors' HIP model, a model property rather than an empirical fact. The gold-standard validation (Ram & Rizoiu 2024) is a self-citation but uses human pairwise judgments external to this paper's fitted values, so it is not counted as circular on its own; however, the same body of prior work is repeatedly used to prop up the causal assumptions. Overall, the counterfactual predictions are well-defined model outputs and the ranking comparison against empirical influence is real empirical content, but the central claim that these differences are causal ATEs is not independently identified and partially reduces, by definition and by self-citation, to the model's own sensitivity.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The method's causal identifiability rests entirely on the three assumptions in Section 3.2, none of which are empirically tested; the counterfactual predictions assume the fitted model generalizes to out-of-distribution inputs; and the data pipeline assumes Google Trends can be treated as a 10-minute-interval exogenous treatment. These are domain assumptions imported into the paper rather than derived or validated.

free parameters (4)
  • alpha (joint loss weight) = 0.5
    Set in Appendix A.2; controls balance between outcome MSE and treatment BCE in the joint loss, affecting model behavior and ATE estimates.
  • beta (temporal/attention decay) = 0.1
    Used in the attention mask and temporal coherence loss; chosen by hand rather than from data.
  • tau_lag (Google Trends lag window) = 24 hours
    Chosen via grid search; determines which external signals are aligned with each engagement observation.
  • w (number of historical windows in treatment intensity) = 6
    Chosen via grid search; controls how many past Google Trends values feed into the intensity model.
assumptions (5)
  • domain assumption Assumption 1 (Consistency): the potential outcome under a pattern of external signals equals the observed outcome when that pattern occurs.
    Needed for ATE identification; the paper argues it holds due to stable platform mechanics, citing prior work but not testing it.
  • domain assumption Assumption 2 (Fully-Mediated Policy Effect): external signals affect engagement only through observable mechanisms, with no hidden pathways.
    This is the no-unobserved-confounding assumption; the paper later lists algorithmic amplification as an unobserved confounder, undermining the assumption.
  • domain assumption Assumption 3 (Temporal Precedence): causes precede effects, so future signals cannot affect current engagement.
    Standard in causal inference; assumed from platform mechanics rather than verified per post.
  • domain assumption The fitted predictive model generalizes to counterfactual inputs outside the training distribution (positivity/overlap).
    Counterfactual scenarios shift treatment intensity and timing; the model's predictions are used as if valid in those regions without overlap checks.
  • domain assumption Google Trends values are a valid, correctly aligned exogenous treatment available at 10-minute intervals.
    The data pipeline is not detailed enough to verify this measurement claim; Google Trends historically does not provide 10-minute resolution for long historical ranges.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Estimating Online Influence Needs Causal Modeling! Counterfactual Analysis of Social Media Engagement." pith.science (2026). https://pith.science/paper/SYBUIBIO

@misc{pith2026250519355,
  author       = {Pith},
  title        = {Pith review of: Estimating Online Influence Needs Causal Modeling! Counterfactual Analysis of Social Media Engagement},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SYBUIBIO}},
  note         = {Machine review of arXiv:2505.19355}
}
read the original abstract

Understanding true influence in social media requires distinguishing correlation from causation--particularly when analyzing misinformation spread. While existing approaches focus on exposure metrics and network structures, they often fail to capture the causal mechanisms by which external temporal signals trigger engagement. We introduce a novel joint treatment-outcome framework that leverages existing sequential models to simultaneously adapt to both policy timing and engagement effects. Our approach adapts causal inference techniques from healthcare to estimate Average Treatment Effects (ATE) within the sequential nature of social media interactions, tackling challenges from external confounding signals. Through our experiments on real-world misinformation and disinformation datasets, we show that our models outperform existing benchmarks by 15--22% in predicting engagement across diverse counterfactual scenarios, including exposure adjustment, timing shifts, and varied intervention durations. Case studies on 492 social media users show our causal effect measure aligns strongly with the gold standard in influence estimation, the expert-based empirical influence.

Figures

Figures reproduced from arXiv: 2505.19355 by the authors.

Figure 1
Figure 1. Visualization of engagement data and queries for social media post [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Engagement trajectory and counterfactual scenarios. Blue dashed lines represent [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Decile Heatmaps (Spearman ρ correlation coefficient, Kendall’s W rank agreement, Concordance Correlation Coefficient CCC): (left) Follower Counts vs. Empirical Influence (ρ = 0.49, W = 0.67, CCC = 0.00), (centre) Causal Effect vs. Empirical Influence (0.57, 0.70, 0.21), (right) Follower Counts vs. Causal Effect (0.32, 0.21, 0.01). Baselines and datasets. There exists no universally accepted ground truth for measurin… view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Relative percentage changes in engagement for three climate change misinforma [PITH_FULL_IMAGE:figures/full_fig_p010_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

56 extracted references · 43 canonical work pages

  1. [1]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION format.date year duplicate empty "emp...

  2. [2]

    Distinguishing influence-based contagion from homophily-driven diffusion in dynamic networks

    Sinan Aral, Lev Muchnik, and Arun Sundararajan. Distinguishing influence-based contagion from homophily-driven diffusion in dynamic networks. Proceedings of the National Academy of Sciences, 106 0 (51): 0 21544--21549, 2009

    work page 2009

  3. [3]

    Everyone's an influencer: quantifying influence on twitter

    Eytan Bakshy, Jake M Hofman, Winter A Mason, and Duncan J Watts. Everyone's an influencer: quantifying influence on twitter. In Proceedings of the fourth ACM international conference on Web search and data mining, pp.\ 65--74, 2011

    work page 2011

  4. [4]

    Exposure to ideologically diverse news and opinion on facebook

    Eytan Bakshy, Solomon Messing, and Lada A Adamic. Exposure to ideologically diverse news and opinion on facebook. Science, 348 0 (6239): 0 1130--1132, 2015

    work page 2015

  5. [5]

    Using social and behavioural science to support covid-19 pandemic response

    Jay J Van Bavel, Katherine Baicker, Paulo S Boggio, Valerio Capraro, Aleksandra Cichocka, Mina Cikara, Molly J Crockett, Alia J Crum, Karen M Douglas, James N Druckman, et al. Using social and behavioural science to support covid-19 pandemic response. Nature human behaviour, 4 0 (5): 0 460--471, 2020

    work page 2020

  6. [6]

    Network dynamics of social influence in the wisdom of crowds

    Joshua Becker, Devon Brackbill, and Damon Centola. Network dynamics of social influence in the wisdom of crowds. Proceedings of the national academy of sciences, 114 0 (26): 0 E5070--E5076, 2017

    work page 2017

  7. [7]

    Estimating counterfactual treatment outcomes over time through adversarially balanced representations

    Ioana Bica, Ahmed M Alaa, James Jordon, and Mihaela van der Schaar. Estimating counterfactual treatment outcomes over time through adversarially balanced representations. In International Conference on Learning Representations, 2020. URL https://openreview.net/forum?id=BJg866NFvB

    work page 2020

  8. [8]

    Opinion market model: stemming far-right opinion spread using positive interventions

    Pio Calderon, Rohit Ram, and Marian-Andrei Rizoiu. Opinion market model: stemming far-right opinion spread using positive interventions. In Proceedings of the international AAAI conference on web and social media, volume 18, pp.\ 177--190, 2024

    work page 2024

Show all 56 references
  1. [9]

    Deephawkes: Bridging the gap between prediction and understanding of information cascades

    Qi Cao, Huawei Shen, Keting Cen, Wentao Ouyang, and Xueqi Cheng. Deephawkes: Bridging the gap between prediction and understanding of information cascades. In Proceedings of the 2017 ACM on Conference on Information and Knowledge Management, pp.\ 1149--1158, 2017

    2017

  2. [10]

    Experimental evidence for tipping points in social convention

    Damon Centola, Joshua Becker, Devon Brackbill, and Andrea Baronchelli. Experimental evidence for tipping points in social convention. Science, 360 0 (6393): 0 1116--1119, 2018

    2018

  3. [11]

    Measuring user influence in twitter: The million follower fallacy

    Meeyoung Cha, Hamed Haddadi, Fabricio Benevenuto, and Krishna Gummadi. Measuring user influence in twitter: The million follower fallacy. In Proceedings of the international AAAI conference on web and social media, volume 4, pp.\ 10--17, 2010

    2010

  4. [12]

    Can cascades be predicted? In Proceedings of the 23rd international conference on World wide web, pp.\ 925--936, 2014

    Justin Cheng, Lada Adamic, P Alex Dow, Jon Michael Kleinberg, and Jure Leskovec. Can cascades be predicted? In Proceedings of the 23rd international conference on World wide web, pp.\ 925--936, 2014

    2014

  5. [13]

    Flashattention: Fast and memory-efficient exact attention with io-awareness

    Tri Dao, Dan Fu, Stefano Ermon, Atri Rudra, and Christopher R \'e . Flashattention: Fast and memory-efficient exact attention with io-awareness. Advances in neural information processing systems, 35: 0 16344--16359, 2022

    2022

  6. [14]

    Social media popularity prediction: A multiple feature fusion approach with deep neural networks

    Keyan Ding, Ronggang Wang, and Shiqi Wang. Social media popularity prediction: A multiple feature fusion approach with deep neural networks. In Proceedings of the 27th ACM International Conference on Multimedia, pp.\ 2682--2686, 2019

    2019

  7. [15]

    Estimating peer effects in networks with peer encouragement designs

    Dean Eckles, Ren \'e F Kizilcec, and Eytan Bakshy. Estimating peer effects in networks with peer encouragement designs. Proceedings of the National Academy of Sciences, 113 0 (27): 0 7316--7322, 2016

    2016

  8. [16]

    Association of mobile phone location data indications of travel and stay-at-home mandates with covid-19 infection rates in the us

    Song Gao, Jinmeng Rao, Yuhao Kang, Yunlei Liang, Jake Kruse, Dorte Dopfer, Ajay K Sethi, Juan Francisco Mandujano Reyes, Brian S Yandell, and Jonathan A Patz. Association of mobile phone location data indications of travel and stay-at-home mandates with covid-19 infection rate...

    2020

  9. [17]

    Mamba: Linear-time sequence modeling with selective state spaces

    Albert Gu and Tri Dao. Mamba: Linear-time sequence modeling with selective state spaces. In First Conference on Language Modeling, 2024. URL https://openreview.net/forum?id=tEYskw1VY2

    2024

  10. [18]

    Hippo: Recurrent memory with optimal polynomial projections

    Albert Gu, Tri Dao, Stefano Ermon, Atri Rudra, and Christopher R \'e . Hippo: Recurrent memory with optimal polynomial projections. Advances in neural information processing systems, 33: 0 1474--1487, 2020

    2020

  11. [19]

    Causal modeling of policy interventions from treatment-outcome sequences

    C a g lar H zl , ST John, Anne Tuulikki Juuti, Tuure Tapani Saarinen, Kirsi Hannele Pietil \"a inen, and Pekka Marttinen. Causal modeling of policy interventions from treatment-outcome sequences. In International Conference on Machine Learning, pp.\ 13050--13084. PMLR, 2023

    2023

  12. [20]

    Parameter-efficient transfer learning for nlp

    Neil Houlsby, Andrei Giurgiu, Stanislaw Jastrzebski, Bruna Morrone, Quentin De Laroussilhe, Andrea Gesmundo, Mona Attariyan, and Sylvain Gelly. Parameter-efficient transfer learning for nlp. In International Conference on Machine Learning, pp.\ 2790--2799. PMLR, 2019

    2019

  13. [21]

    \# HashtagActivism: Networks of race and gender justice

    Sarah J Jackson, Moya Bailey, and Brooke Foucault Welles. \# HashtagActivism: Networks of race and gender justice. Mit Press, 2020

    2020

  14. [22]

    Slipping to the E xtreme: A M ixed M ethod to E xplain H ow E xtreme O pinions I nfiltrate O nline D iscussions

    Quyu Kong, Emily Booth, Francesco Bailo, Amelia Johns, and Marian-Andrei Rizoiu. Slipping to the E xtreme: A M ixed M ethod to E xplain H ow E xtreme O pinions I nfiltrate O nline D iscussions. In Proceedings of the International AAAI Conference on Web and Social Media, volume...

    2022

  15. [23]

    Interval-censored transformer hawkes: Detecting information operations using the reaction of social systems

    Quyu Kong, Pio Calderon, Rohit Ram, Olga Boichak, and Marian-Andrei Rizoiu. Interval-censored transformer hawkes: Detecting information operations using the reaction of social systems. In Proceedings of the ACM web conference 2023, pp.\ 1813--1821, 2023

    2023

  16. [24]

    What is twitter, a social network or a news media? In Proceedings of the 19th international conference on World wide web, pp.\ 591--600, 2010

    Haewoon Kwak, Changhyun Lee, Hosung Park, and Sue Moon. What is twitter, a social network or a news media? In Proceedings of the 19th international conference on World wide web, pp.\ 591--600, 2010

    2010

  17. [25]

    Deepcas: An end-to-end predictor of information cascades

    Cheng Li, Jiaqi Ma, Xiaoxiao Guo, and Qiaozhu Mei. Deepcas: An end-to-end predictor of information cascades. In Proceedings of the 26th international conference on World Wide Web, pp.\ 577--586, 2017

    2017

  18. [26]

    Forecasting treatment responses over time using recurrent marginal structural networks

    Bryan Lim. Forecasting treatment responses over time using recurrent marginal structural networks. Advances in neural information processing systems, 31, 2018

    2018

  19. [27]

    Statistical modeling of causal effects in continuous time

    Judith J Lok. Statistical modeling of causal effects in continuous time. The Annals of Statistics, pp.\ 1464--1507, 2008

    2008

  20. [28]

    Continuous-time graph learning for cascade popularity prediction

    Xiaodong Lu, Shuo Ji, Le Yu, Leilei Sun, Bowen Du, and Tongyu Zhu. Continuous-time graph learning for cascade popularity prediction. In Proceedings of the Thirty-Second International Joint Conference on Artificial Intelligence, pp.\ 2224--2232, 2023

    2023

  21. [29]

    Social influence bias: A randomized experiment

    Lev Muchnik, Sinan Aral, and Sean J Taylor. Social influence bias: A randomized experiment. Science, 341 0 (6146): 0 647--651, 2013

    2013

  22. [30]

    Causality

    Judea Pearl. Causality. Cambridge university press, 2009

    2009

  23. [31]

    Elements of causal inference: foundations and learning algorithms

    Jonas Peters, Dominik Janzing, and Bernhard Sch \"o lkopf. Elements of causal inference: foundations and learning algorithms. The MIT Press, 2017

    2017

  24. [32]

    Deepinf: Social influence prediction with deep learning

    Jiezhong Qiu, Jian Tang, Hao Ma, Yuxiao Dong, Kuansan Wang, and Jie Tang. Deepinf: Social influence prediction with deep learning. In Proceedings of the 24th ACM SIGKDD international conference on knowledge discovery & data mining, pp.\ 2110--2119, 2018

    2018

  25. [33]

    Empirically measuring online social influence

    Rohit Ram and Marian-Andrei Rizoiu. Empirically measuring online social influence. EPJ Data Science, 13 0 (1): 0 53, 2024

    2024

  26. [34]

    Online popularity under promotion: Viral potential, forecasting, and the economics of time

    Marian-Andrei Rizoiu and Lexing Xie. Online popularity under promotion: Viral potential, forecasting, and the economics of time. In International AAAI Conference on Web and Social Media (ICWSM '17), pp.\ 182--191, 2017 a . URL https://aaai.org/ocs/index.php/ICWSM/ICWSM17/paper...

    2017 arXiv

  27. [35]

    Online popularity under promotion: Viral potential, forecasting, and the economics of time

    Marian-Andrei Rizoiu and Lexing Xie Xie. Online popularity under promotion: Viral potential, forecasting, and the economics of time. In Proceedings of the International AAAI Conference on Web and Social Media, volume 11, pp.\ 182--191, 2017 b

    2017

  28. [36]

    Expecting to be hip: Hawkes intensity processes for social media popularity

    Marian-Andrei Rizoiu, Lexing Xie, Scott Sanner, Manuel Cebrian, Honglin Yu, and Pascal Van Hentenryck. Expecting to be hip: Hawkes intensity processes for social media popularity. In Proceedings of the 26th international conference on world wide web, pp.\ 735--744, 2017

    2017

  29. [37]

    Interval-censored hawkes processes

    Marian-Andrei Rizoiu, Alexander Soen, Shidi Li, Pio Calderon, Leanne J Dong, Aditya Krishna Menon, and Lexing Xie. Interval-censored hawkes processes. Journal of Machine Learning Research, 23 0 (338): 0 1--84, 2022

    2022

  30. [38]

    A new approach to causal inference in mortality studies with a sustained exposure period—application to control of the healthy worker survivor effect

    James Robins. A new approach to causal inference in mortality studies with a sustained exposure period—application to control of the healthy worker survivor effect. Mathematical modelling, 7 0 (9-12): 0 1393--1512, 1986

    1986

  31. [39]

    Marginal structural models and causal inference in epidemiology, 2000

    James M Robins, Miguel Angel Hernan, and Babette Brumback. Marginal structural models and causal inference in epidemiology, 2000

    2000

  32. [40]

    Social acceleration: A new theory of modernity

    Hartmut Rosa. Social acceleration: A new theory of modernity. Columbia University Press, 2013

    2013

  33. [41]

    Reliable decision support using counterfactual models

    Peter Schulam and Suchi Saria. Reliable decision support using counterfactual models. Advances in neural information processing systems, 30, 2017

    2017

  34. [42]

    Treatment-response models for counterfactual reasoning with continuous-time, continuous-valued interventions

    Hossein Soleimani, Adarsh Subbaswamy, and Suchi Saria. Treatment-response models for counterfactual reasoning with continuous-time, continuous-valued interventions. In 33rd Conference on Uncertainty in Artificial Intelligence, UAI 2017, 2017

    2017

  35. [43]

    Before it's too late: A state space model for the early prediction of misinformation and disinformation engagement

    Lin Tian, Emily Booth, Francesco Bailo, Julian Droogan, and Marian-Andrei Rizoiu. Before it's too late: A state space model for the early prediction of misinformation and disinformation engagement. arXiv preprint arXiv:2502.04655, 2025

    2025 arXiv

  36. [44]

    Attention is all you need

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, ukasz Kaiser, and Illia Polosukhin. Attention is all you need. In Advances in Neural Information Processing Systems, volume 30. Curran Associates, Inc., 2017. URL https://proceedings.neurip...

    2017

  37. [45]

    Topological recurrent neural network for diffusion prediction

    Jia Wang, Vincent W Zheng, Zemin Liu, and Kevin Chen-Chuan Chang. Topological recurrent neural network for diffusion prediction. In 2017 IEEE international conference on data mining (ICDM), pp.\ 475--484. IEEE, 2017 a

    2017

  38. [46]

    Cascade dynamics modeling with attention-based recurrent neural network

    Yongqing Wang, Huawei Shen, Shenghua Liu, Jinhua Gao, and Xueqi Cheng. Cascade dynamics modeling with attention-based recurrent neural network. In IJCAI, volume 17, pp.\ 2985--2991, 2017 b

    2017

  39. [47]

    Measuring the news and its impact on democracy

    Duncan J Watts, David M Rothschild, and Markus Mobius. Measuring the news and its impact on democracy. Proceedings of the National Academy of Sciences, 118 0 (15): 0 e1912443118, 2021

    2021

  40. [48]

    Autoformer: Decomposition transformers with auto-correlation for long-term series forecasting

    Haixu Wu, Jiehui Xu, Jianmin Wang, and Mingsheng Long. Autoformer: Decomposition transformers with auto-correlation for long-term series forecasting. Advances in neural information processing systems, 34: 0 22419--22430, 2021

    2021

  41. [49]

    Attitudinal effects of mere exposure

    Robert B Zajonc. Attitudinal effects of mere exposure. Journal of personality and social psychology, 9 0 (2p2): 0 1, 1968

    1968

  42. [50]

    Seismic: A self-exciting point process model for predicting tweet popularity

    Qingyuan Zhao, Murat A Erdogdu, Hera Y He, Anand Rajaraman, and Jure Leskovec. Seismic: A self-exciting point process model for predicting tweet popularity. In Proceedings of the 21th ACM SIGKDD international conference on knowledge discovery and data mining, pp.\ 1513--1522, 2015

    2015

  43. [51]

    A survey of information cascade analysis: Models, predictions, and recent advances

    Fan Zhou, Xovee Xu, Goce Trajcevski, and Kunpeng Zhang. A survey of information cascade analysis: Models, predictions, and recent advances. ACM Computing Surveys (CSUR), 54 0 (2): 0 1--36, 2021 a

    2021

  44. [52]

    Informer: Beyond efficient transformer for long sequence time-series forecasting

    Haoyi Zhou, Shanghang Zhang, Jieqi Peng, Shuai Zhang, Jianxin Li, Hui Xiong, and Wancai Zhang. Informer: Beyond efficient transformer for long sequence time-series forecasting. In Proceedings of the AAAI conference on artificial intelligence, volume 35, pp.\ 11106--11115, 2021 b

    2021

  45. [53]

    Transformer hawkes process

    Simiao Zuo, Haoming Jiang, Zichong Li, Tuo Zhao, and Hongyuan Zha. Transformer hawkes process. In International conference on machine learning, pp.\ 11692--11702. PMLR, 2020

    2020

  46. [54]

    @esa (Ref

    \@ifxundefined[1] #1\@undefined \@firstoftwo \@secondoftwo \@ifnum[1] #1 \@firstoftwo \@secondoftwo \@ifx[1] #1 \@firstoftwo \@secondoftwo [2] @ #1 \@temptokena #2 #1 @ \@temptokena \@ifclassloaded agu2001 natbib The agu2001 class already includes natbib coding, so you should ...

  47. [55]

    \@lbibitem[] @bibitem@first@sw\@secondoftwo \@lbibitem[#1]#2 \@extra@b@citeb \@ifundefined br@#2\@extra@b@citeb \@namedef br@#2 \@nameuse br@#2\@extra@b@citeb \@ifundefined b@#2\@extra@b@citeb @num @parse #2 @tmp #1 NAT@b@open@#2 NAT@b@shut@#2 \@ifnum @merge>\@ne @bibitem@firs...

  48. [56]

    @open @close @open @close and [1] URL: #1 \@ifundefined chapter * \@mkboth \@ifxundefined @sectionbib * \@mkboth * \@mkboth\@gobbletwo \@ifclassloaded amsart * \@ifclassloaded amsbook * \@ifxundefined @heading @heading NAT@ctr thebibliography [1] @ \@biblabel @NAT@ctr \@bibset...

    2020

Pith tools

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