# chronofact_timelinebased_temporal_fact_verification__a85949d5.pdf Chrono Fact: Timeline-based Temporal Fact Verification Anab Maulana Barik1 , Wynne Hsu1,2 and Mong Li Lee1,3 1School of Computing, National University of Singapore, Singapore 2Institute of Data Science, National University of Singapore, Singapore 3Centre for Trusted Internet and Community, National University of Singapore, Singapore anabmaulana@u.nus.edu, {whsu,leeml}@comp.nus.edu.sg Temporal claims, often riddled with inaccuracies, are a significant challenge in the digital misinformation landscape. Fact-checking systems that can accurately verify such claims are crucial for combating misinformation. Current systems struggle with the complexities of evaluating the accuracy of these claims, especially when they include multiple, overlapping, or recurring events. We introduce a novel timeline-based fact verification framework that identify events from both claim and evidence and organize them into their respective chronological timelines. The framework systematically examines the relationships between the events in both claim and evidence to predict the veracity of each claim event and their chronological accuracy. This allows us to accurately determine the overall veracity of the claim. We also introduce a new dataset of complex temporal claims involving timeline-based reasoning for the training and evaluation of our proposed framework. Experimental results demonstrate the effectiveness of our approach in handling the intricacies of temporal claim verification. 1 Introduction The spread of false information has reached alarming levels, undermining social trust significantly. One prominent category of false information involves temporal claims, which are statements that include time-specific elements, either explicitly (e.g., in 1953 ) or implicitly (e.g., before [another event] ). The complexity of verifying these claims escalates with the number of events mentioned. Effective verification must assess not only the veracity of each event within its temporal context but also understand the relationships between these events, especially their chronological order. Existing research has largely focused on verifying the veracity of individual events within claims, and overlook the chronological order of these events [Barik et al., 2024; Qudus et al., 2023]. This oversight can undermine the effectiveness of fact-checking systems, particularly when dealing with complex narratives where the sequence of events is crucial for determining the truth. Understanding the chronological order of events is hindered by the absence of the explicit (a) Claim with multiple events (b) Claim with recurring event Figure 1: Illustration of complex temporal claim verification. temporal cues. The complexity increases when events overlap or recur, complicating the task of establishing a clear and accurate timeline for the verification process. Example 1. Figure 1(a) shows a claim involving four events, with evidence relevant to each event indicated by matching color boxes. While each event appears to be supported by some evidence when analyzed independently, closer scrutiny of the timeline reveals discrepancies. For instance, evidence indicates that negotiations between Ukraine and Russia took place in Feb 2022, before the bombardments in March 2023. Therefore, the claim event Russia and Ukraine participated in a series of ceasefire negotiations , suggested by the temporal cues to have occurred in March 2023, is not actually supported by the evidence. Existing works that analyze individ- Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence (IJCAI-25) Figure 2: Overview of Chrono Fact Framework. ual events without considering the timeline would incorrectly conclude that all the events in the claim are supported, and deem the claim true. This shows the importance of chronological order of events in claim verification. Example 2. Figure 1(b) shows a claim event Israel attacks Hamas , with evidence indicating it occurs at three different times: Nov 2023, Jan 2024, and Aug 2024. Analysis of the timeline shows that the occurrence in Aug 2024 aligns with the order of events in the claim, thereby concluding that the claim is supported. Previous studies that have overlooked the timeline of events might match the claim event to earlier occurrence in Nov 2023, or Jan 2024, leading to incorrect conclusion that this claim is refuted. This highlights the need to align claim and evidence events with their chronological order for accurate claim verification. To overcome the limitations in existing temporal fact verification methods, we introduce Chrono Fact, a framework that systematically identify events from both claim and evidence to construct a coherent timeline for temporal fact verification. Chrono Fact examines the relationships between claim and evidence events at three levels: event-level, tokenlevel, and time-level and learns a model to predict the veracity of each claim event and their chronological accuracy. We develop a new dataset called Chrono Claims for timeline-based fact verification. This dataset encompasses complex claims involving multiple, related events that unfold over time with both implicit and explicit temporal expressions. These features are lacking in current datasets like T-FEVER, which mainly consist of single-event claims, and T-FEVEROUS, which, despite its more complex claims, often has events that are not chronologically related. 2 Related Work There has been a stream of research on evidence-based claim verification to assess whether the evidence sentences support or refute the claim [Stammbach and Neumann, 2019; Soleimani et al., 2020]. Techniques like GEAR [Zhou et al., 2019], KGAT [Liu et al., 2020], and DREAM [Zhong et al., 2020] model claim and evidence into a graph using Graph Attention Network to facilitate information propagation of information between them. CGAT [Barik et al., 2022] enhances this by incorporating commonsense knowledge from Concept Net to enrich the contextual representation. These works do not consider temporal information. Only a few studies have incorporated temporal information in the claim verification process. NMT [Mori et al., 2022] verifies economic claims against time series data in tabular format by translating the claims into Datalog rules to query the tabular evidence. Temporal FC [Qudus et al., 2023] focuses on verifying claims represented as tuples against a knowledge graph. It utilizes temporal graph embeddings to determine the validity timing of underlying triples. However, these works are limited to structured data, hindering its applicability to natural language claims and evidence. [Allein et al., 2021] capture the temporal relevance of evidence through a re-ranking process based on the proximity of publication dates between the evidence and the claim. ITR [Allein et al., 2023] extends this by assigning publication dates of claims and evidence to fixed-size time buckets for temporal reasoning. However, it ignores implicit temporal cues or chronological order in the claim and evidence. TACV [Barik et al., 2024] decomposes claim and evidence sentences into events and employ temporal-aware representation encoder to retrieve evidence that are both semantically and temporally related to the claim. It utilizes GPT for temporal reasoning to verify individual claim events. Unlike TACV, our approach also evaluates the chronological timeline accuracy between the claim and evidence, providing a more thorough verification of complex claims. A related area of research, question answering (QA), has also begun to incorporate temporal information. FAITH [Jia et al., 2024] focuses on implicit temporal reasoning for the temporal question answering task by generating intermediate questions with explicit temporal information. This method uses heterogeneous sources to improve the completeness of evidence retrieval and employs a general-purpose QA system to respond to the question. While it is common to utilize QA in the claim verification process [Pan et al., 2023], they do not consider the chronological order of events. In contrast, our work focuses on verifying complex claims that require timeline-based reasoning. 3 Proposed Framework Chrono Fact follows the typical automated claim verification which involves collecting relevant evidence from credible sources and assessing the claim s veracity based on the evidence. Figure 2 shows the key modules in the framework. Event Extractor. Given a claim, we first employ the GENRE sequence-to-sequence entity linking model [De Cao et al., 2020] to retrieve relevant documents from Wikipedia and extract all evidence sentences. Then we utilize Semantic Role Labelling from Allen NLP [Allen, 1983] to extract events from both the claim and evidence. Each event has its core information and temporal argument. Finally, we score the evidence with the events extracted from the claim using an event representation encoder similar to [Barik et al., 2024]. Event Encoder. We tokenize each event and pass each token i to the flan-T5 model to obtain the corresponding token representations Hi. For tokens that represent date, we apply mean pooling, followed by positional encoding [Vaswani et al., 2017] where the position corresponds to the distance between the event date and the earliest date found in the claim and evidence events. The final representation of the event is given by where d is the number of tokens in the event, HCLS is the event-level representation obtained by the average pooling of Hj, 1 j d, and H1...Hd are the token-level representations. Multi-level Attention Encoder. We process each pair of claim event representation ci and evidence event representa- Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence (IJCAI-25) tion ej through a multi-level attention module to determine the relevance of evidence events for each claim event. This involves calculating attention scores at three levels: token-level attention score αij: is the average of cosine similarities between all pairs of tokens in ci and ej. event-level attention score βij: is the cosine similarity between the HCLS representations of ci and ej. time-level attention score γij: is the cosine similarity between the mean pooled date representations in ci and ej. The final multi-level attention score between ci and ej, denoted as ωij, is the average of the event-level, token-level, and time-level attention scores. These computed attention scores are then employed to predict the label of each event in the claim, assess the accuracy of the claim s chronological order based on the evidence timeline, and evaluate the claim s overall veracity. Note that we considered dynamically learn the weights of each attention level. However, our experiment indicated that this approach did not improve performance, leading us to adopt the average weighting method instead. Claim Event Classifier. This module predicts the label of claim event ci using the top-k evidence events E E with the highest final attention scores. Let Hci CLS and Hej CLS be the representation of the CLS token for ci and evidence event ej E respectively. We concatenate the representations weighted by the attention scores to obtain uci as follows: uci = Hci CLS ωi1He1 CLS ωik Hek CLS This is then fed to two fully connected layers followed by softmax to obtain the probability distribution zci: zci = softmax(FC2(Re LU(FC1(uci)))) where zci[0], zci[1], and zci[2] are the probabilities of the labels SUP , REF , and NEI respectively. The label yci for the claim event ci is assigned based on the highest probability among these. Chronological Order Classifier. This module predicts the accuracy of the chronological order of a claim C using the timeline of evidence events. Given n events in C, the relevance of each evidence event to C is given by: rej = tanh Xn i=1 ωi,j (1) We use tanh for its bounded output range of [ 1, 1], which aligns with our interpretation of relevance as a continuous spectrum ranging from negative to positive. We sort the evidence events based on their relevance to the claim and obtain the top-k events with the highest score rej. These top-k evidence events, along with the claim events, are then input into GPT to be reordered according to their chronological sequence. Given the limited number k of events, this reordering process requires minimal time and resources, making it practical for real-world applications. The reordered sequence of claim events, denoted as seq C = Hc1 CLS Hcn CLS is passed to a Bi-LSTM to capture and embed the chronological order into the model. Similarly, the reordered sequence of evidence events, weighted by their relevance scores is given by seq E = re1He1 CLS rem Hem CLS This sequence is passed to a second Bi-LSTM. The outputs from the two Bi-LSTM, denoted as o C for claim events and o E for evidence events, are then fed into two fully connected layers followed by softmax to obtain the distribution zo: zo = softmax(FC4(Re LU(FC3([o C o E])))) where zo[0] is the probability that the chronological order of the claim events is supported by that of the evidence events, and zo[1] is the probability that the chronological order refutes the claim s timeline. The output of the chronological order classifier yo is the label with the highest probability. Claim Classifier. To predict the overall veracity of the claim, we concatenate seq C and seq E, along with the distributions zc1 zcn, zo, and pass this vector to two fully connected layers followed by a softmax function to obtain the probability distribution z that the claim is SUP, REF or NEI. The label with the highest probability is depicted as y. 3.1 Model Training We train the model using two losses Lcross and Lsoft. The first loss Lcross is defined as follows: i=1 F(gci, zci) + F(go, zo) + F(g, z) (2) where F(.) is a cross-entropy function, gci, go, and g are the ground-truth labels for claim event ci, chronological order, and overall claim label respectively. The second loss ensures the consistency between the overall claim label, claim event labels, and chronological accuracy. In particular, we apply a set of logic rules based on the outcomes of the individual claim events and their chronological alignment with the evidence. Specifically, a claim is deemed supported if all its associated claim events are supported and their chronological sequence matches that of the evidence events. This is expressed using first-order logic: yc1 ycn yo = y This logical expression states that the overall veracity y of a claim is SUP if and only if each claim event yci is supported and the chronological order yo is consistent with the evidence. On the other hand, if any one of the claim event is refuted, or the chronological order does not align, then y is REF. Otherwise, y is NEI. We leverage G odel t-norm to soften the hard reasoning rules in the claim classifier, and obtain the differentiable distribution zsoft: zsoft[0] = min(zc1[0], , zcn[0], zo[0]) zsoft[1] = max(zc1[1], , zcn[1], zo[1]) zsoft[2] = 1 zsoft[0] zsoft[1] (3) With this, we define Lsoft which ensures the consistency of the overall claim label with the claim events and their chronological order as follows: Lsoft = DKL(z||zsoft) (4) where DKL is Kullback-Leibler divergence, which measures the difference between the predicted overall claim label distribution and the distribution derived from the soft logic. Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence (IJCAI-25) Temporal Expression Train Set Validation Set Test Set Support Refute Support Refute Support Refute 3 events 3,561 2,954 355 318 397 315 4 events 3,574 3,376 354 282 359 262 5 events 3,597 3,084 266 198 323 218 3 events 3,804 3,028 355 318 390 313 4 events 3,453 3,306 355 278 361 263 5 events 3,473 3,039 266 199 321 213 Temporal Category Overlapping events 10,360 9,297 918 807 985 714 Recurring events 8,914 6,507 729 376 865 361 Table 1: Characteristics of Chrono Claims dataset. The final loss function combines the soft logic loss Lsoft with standard cross-entropy loss Lcross, enabling the model to be trained both with supervision and structured constraints, with mu as the hyperparameter: L = (1 µ)Lcross + µLsoft (5) 4 Chrono Claims Dataset We introduce a new benchmark dataset called Chrono Claims that is designed for enhancing the accuracy and complexity of timeline-based fact verification. Utilizing the November 2022 Wikidata snapshot [Vrandeˇci c and Kr otzsch, 2014], we preprocess and extract facts in the format . To construct an evidence timeline, we organize all the facts that have the same subject in a chronological order. Then we randomly select N facts from this timeline and use GPT to transform these facts into coherent sentences. For generating sentences with implicit temporal information, we omit the time start and time end from the fact, and prompt GPT to craft sentences that subtly embed the temporal context. Then we use a cloze-style template1 to synthesize the sentences into a claim that preserves the chronological order. To generate claims that contradict the evidence timeline, we rearrange the order of these sentences. Finally, we use GPT to refine and rephrase the synthetic claim to make it more natural and fluent. Each claim is labeled as either SUP or REF, depending on whether it aligns or conflicts with the timeline. We evaluate the quality of the Chrono Claims dataset by sampling 500 claims, ensuring equal distribution across different temporal expressions and event complexities (16% each type), and temporal categories (50% each category) to achieve representative coverage. Two human annotators are tasked with determining the labels of the generated claims based on the ground truth evidence timeline. The agreement rates are high, with 96.8% and 97% of the labels assigned by the annotators matching the labels of the generated claims. We analyze the claims where the generated labels are different from the annotators, and discover that most of the discrepancies are due to errors in rephrasing For example, a claim Nasrallah Peter Sfeir worked as a Catholic priest, and then Nasrallah Peter Sfeir was educated at Saint Joseph University, was rephrased to Nasrallah Peter Sfeir pursued his education at Saint Joseph University before becoming a Catholic priest, which altered the original chronological sequence. To rectify this, we perform a second verification 1 and then and then step using GPT to ensure that the semantic meanings of the original and rephrased claims remain consistent. Any claims where the meaning has been altered are discarded. In total, we generated 40,249, 3,544 and 3,735 claims for the training, validation and test sets. Table 1 shows the detailed statistics of the Chrono Claims dataset. 5 Performance Study Datasets. Besides the Chrono Claims dataset, we also use the T-FEVER and T-FEVEROUS datasets [Barik et al., 2024]. These datasets are derived from the benchmark fact verification datasets FEVER [Thorne et al., 2018a], FEVER2.0 [Thorne et al., 2018b] and FEVEROUS [Aly et al., 2021] respectively such that the claims in T-FEVER and T-FEVEROUS contain temporal expressions. Each claim involves a maximum of 3 events and is labeled as SUP, REF and NEI. We also evaluate our method on the T-Quan Temp, a subset of the Quan Temp [Venktesh et al., 2024] dataset focusing on real-world claims with temporal aspects. TQuan Temp comprises of temporal claims containing temporal expressions or tagged as as temporal aspects by Quan Temp dataset. Table 2 shows the characteristics of T-FEVER, T-FEVEROUS and T-Quan Temp datasets. We use 80% of the data for training and 20% for testing. While T-Quan Temp comes with its own set of evidence, we rely on different knowledge sources for the others: Wikipedia for T-FEVER and T-FEVEROUS, and Wikidata for Chrono Claims. Dataset Support Refute NEI T-FEVER 11,799 9,292 3,975 T-FEVEROUS 33,357 28,959 1,266 T-Quan Temp 1,261 3,470 1,349 Table 2: Dataset characteristics. Implementation Details. We implement the Chrono Fact framework using Hugging Face Transformers Library with Py Torch. The Event Encoder use flan-T5 base [Chung et al., 2024] with a hidden size of 768. In the Multi-level Attention Encoder, the token-level, event-level, and time-level representations pass through a linear layer of dimension 768 to calculate attention scores. The hidden size of the fully connected layers is set to 192. The Chronological Order Classifier uses two layers of Bi-LSTM, each with a hidden size of 768, and the fully connected layers have a hidden size of 192, matching those in the Claim Classifier. We train the model using Adafactor with a batch size of 8 and a learning rate of 5e-5 for 5 epochs on each dataset. Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence (IJCAI-25) (a) Effect on µ (b) Effect on k Figure 3: Sensitivity Experiments on Chrono Fact. Chrono Fact is trained to predict REF and SUP labels on Chrono Claims, as the dataset does not have NEI labels. For T-FEVER and T-FEVEROUS, the model is trained to predict SUP, REF, and NEI. We report the macro F1 score of the best performing model on the test set. 5.1 Sensitivity Experiments We first conduct sensitivity experiments to obtain the optimal value of the parameter µ in Eqn 5. Figure 3(a) shows the label accuracy as we vary the value of µ from 0.1 to 0.9. The performance improves when µ increases from 0.1 to 0.3, indicating that the model benefits from incorporating Lsoft. The best performance is achieved when µ = 0.3 across all datasets and we use this value for the rest of the experiments. We also vary the number of top-k evidence events in the classifier module to examine its effect on Chrono Fact s performance. Figure 3(b) shows that the optimal performance for T-FEVER, T-FEVEROUS, and Chrono Claims was achieved when k is 3, 5, and 7 respectively, and we use these values in our experiments. Note that Chrono Claims generally achieves higher performance because it has only two class labels (SUPPORTS, REFUTES), whereas T-FEVER and TFEVEROUS include an additional label ( NOT ENOUGH INFO ). 5.2 Comparative Experiments We compare Chrono Fact with the following state-of-theart evidence-based fact verification baselines: KGAT [Liu et al., 2020]. This uses a transformer to obtain claim-sentence representations and a graph attention network to aggregate the evidence for claim verification. CGAT [Barik et al., 2022]. This method incorporates external knowledge from Concept Net to enrich the contextual representations of claim and evidence sentences. It then employs graph attention networks to propagate the information among the evidence sentences to verify the claim veracity. ITR [Allein et al., 2023]. This also employs a transformer to obtain the claim and evidence representations which are augmented with the publication dates of the claim and evidence for temporal reasoning. FAITH [Jia et al., 2024]. We adapt this temporal QA model for temporal claim verification by using GPT to generate relevant temporal questions for FAITH, then prompting GPT to verify the claim based on FAITH s generated answers. GPT-4o in a zero-shot setting. Given the claim and retrieved evidence using TACV retrieval method, we prompt GPT-4o to predict the claim s label. TACV [Barik et al., 2024]. This is an end-to-end solution for temporal claim verification that considers the temporal information in claims to obtain relevant evidence sentences and uses LLM for temporal reasoning. Table 3 shows the macro F1 score of the various methods on the Chrono Claims dataset, with our proposed Chrono Fact achieving the highest score and significantly outperforming the baseline models2. Further analysis in terms of the number of events shows that Chrono Fact is better in managing complex claims with multiple events. This is evident in the smaller decline in the performance as the number of events increases, compared to other baselines. We also analyze Chrono Fact s ability to handle complex temporal event types where the timelines of different events may overlap or the events may recur at multiple time points. Once again, there is a significant gap in the macro F1 score compared to other methods, demonstrating Chrono Fact s superior performance in managing such complexities. Table 4 shows the performance on T-FEVER and TFEVEROUS datasets, where Chrono Fact is the best performer on both datasets. Similar trend is observed when comparing the performance in terms of the number of events. Table 4 also shows Chrono Fact s robustness on the realworld T-Quan Temp, despite noisy date information. Error Analysis. We conduct an error analysis on 50 randomly incorrectly predicted claims on Chrono Claim dataset. We find that 82% of these errors are due to claims containing implicit temporal information. For example, the claim Greg Clark first held the position of Minister of State for Decentralisation, thereafter became the Secretary of State for Business and Trade, and later served as the Secretary of State for Levelling Up, Housing and Communities. lacks explicit dates or clear temporal markers. This ambiguity poses a challenge for the model to accurately encode and reason about the sequence of events. The remaining 18% of the errors comes from claims with events that have multiple dates and varying levels of granularity, such as Nicola Fratoianni joined the Movement for the Left from 2009 until October 22, 2010 . The complexity of this claims hinders the model s ability to encode the temporal information. 5.3 Ablation Studies We examine the effectiveness of the key modules in Chrono Fact by implementing the following variants: Chrono Fact without multi-level attention encoder module. In this variant, we set the final attention scores between each claim event ci and evidence event ej to 1. Chrono Fact without claim event classifier. Here, the input to the claim classifier is the concatenation of seq C, seq E, and the distribution of chronological order zo. Chrono Fact without chronological order classifier. This variant does not assess the consistency of the chronological order of the claim events with that of the evidence events. Therefore, the input to the Claim Classifier is the concatenation of seq C, seq E, and the probability distribution of the claim events zc1 zcn. 2Results of micro F1 scores are provided in https://arxiv.org/pdf/ 2410.14964 Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence (IJCAI-25) Method Overall Number of Events Event Types 3 events 4 events 5 events Overlapping Recurring KGAT 50.56 3.52 52.61 3.20 49.33 4.06 47.17 5.57 50.14 3.88 47.43 1.68 CGAT 56.69 0.36 58.11 1.20 55.98 0.57 54.75 0.58 56.56 0.22 56.95 0.97 ITR 58.34 6.44 63.86 3.49 61.52 0.27 51.15 2.41 57.87 5.25 55.50 6.31 FAITH 60.84 0.11 61.64 0.89 61.48 0.30 59.03 1.31 61.03 0.21 52.42 0.16 GPT4o 61.93 0.30 70.06 0.33 57.65 0.54 55.51 0.48 61.86 0.22 55.08 0.18 TACV 63.42 0.60 67.45 0.87 63.02 0.78 61.81 0.56 62.37 0.69 65.84 5.04 Chrono Fact 85.49 0.53 86.33 0.46 85.55 0.31 84.86 0.43 84.86 0.76 80.89 0.55 Table 3: Comparison of macro F1 on Chrono Claims. T-FEVER T-FEVEROUS T-Quan Temp Overall Number of Events Overall Number of Events Overall 1 event 2 events 1 event 2 event 3 event KGAT 40.66 1.04 40.82 1.21 37.85 2.40 17.66 1.21 21.95 3.03 17.85 1.39 16.40 0.82 39.82 0.46 CGAT 42.31 2.47 42.47 2.61 39.20 0.85 19.62 1.92 23.78 1.66 19.30 1.85 18.54 1.95 42.01 0.25 ITR 45.21 3.86 45.62 3.89 36.77 2.36 27.62 2.99 30.19 4.50 27.64 3.39 26.99 2.55 44.90 1.27 FAITH 49.03 0.79 49.42 0.57 42.65 4.06 38.63 0.02 41.89 0.15 40.17 0.27 37.75 0.15 49.65 0.24 GPT4o 53.77 0.23 53.54 0.30 57.19 1.29 43.54 0.19 44.28 0.48 44.11 0.22 43.00 0.27 53.12 0.21 TACV 49.86 0.40 50.25 0.41 43.64 0.01 39.84 0.60 43.72 0.84 39.78 0.93 37.72 0.82 49.69 0.47 Chrono Fact 56.29 1.50 56.14 1.41 57.34 2.39 47.78 0.98 48.27 1.62 48.23 2.41 47.88 2.21 65.67 0.20 Table 4: Comparison of macro F1 on T-FEVER, T-FEVEROUS, and T-Quan Temp. Variants Datasets Chrono Claims T-FEVER T-FEVEROUS T-Quan Temp w/o multi-level attention encoder 84.65 0.30 53.60 0.91 42.01 0.55 63.53 1.51 w/o claim event classifier 83.72 0.57 54.55 1.10 41.24 2.28 63.95 0.98 w/o chronological order classifier 76.88 2.14 55.31 0.32 40.75 1.14 64.35 0.43 Chrono Fact 85.49 0.53 56.29 1.50 47.78 0.98 65.67 0.20 Table 5: Macro F1 score of ablation studies. Table 5 shows that the largest performance drop occurs when when we exclude the chronological order classifier, emphasizing that predicting chronological order enhances the model prediction. This effect is particularly evident in the Chrono Claims dataset, which is specifically designed to test the model s ability to reason over the chronological order of events. Its effect is less noticeable on the T-FEVER dataset as it mainly consists of single-event claims. The next largest drop in macro F1 score is when we exclude the claim event classifier, highlighting the importance of predicting individual claim events for the overall claim prediction. Excluding the multi-level attention encoder leads to a decrease in macro F1 score across all datasets, indicating the role of considering the relevance of evidence events in claim verification. We conduct a manual analysis of the attention scores associated with the relevant evidence events for 50 claim events. For each claim event, we examine whether the evidence event with the highest event-level and token-level attention scores is semantically related, and whether the evidence event with the highest time-level attention score is temporally related. Our findings indicate that, in every case, the evidence event with the highest score is indeed semantically or temporally relevant to the corresponding claim event. 6 Case Studies Finally, we present case studies to illustrate the importance of considering the chronological order in verifying temporal claims. Table 6 shows a T-FEVEROUS claim with three events involving Yossi Yona: studying for a Ph D , become a Professor , and join the left camp . The temporal context provided by the word before is lost between the events became a Professor and joined the Left Camp , resulting in each individual claim event being supported by the retrieved evidence. However, the chronological order of the claim events is incorrect because evidence shows that joined the left camp before became a Professor . Chrono Fact addresses this issue by inferring the missing temporal relationships from the timeline of events, and correctly predicts the label as REF. In contrast, TACV evaluates the claim events independently and mistakenly predicts the claim as SUP. Table 7 shows a claim from Chrono Claim dataset. The claim has five events, where the first event Davie was a member of Dundee United F.C overlaps with the second event Davie joined Arbroath F.C. . TACV does not accommodate the possibility of overlapping events and assumes that join implies a full transfer or departure, and conclude that the evidence refutes the claim event c2. In contrast, Chrono Fact can handle overlapping events and correctly conclude that c2 is supported by the evidence. Table 8 shows another claim which involves a recurring event where Andonov was a member of Botev Polediv from 2002 to 2006, left, and rejoined in 2009. TACV fails to recognize this recurrence, mistakenly using evidence of Andonov s initial membership period (2002-2006) to refute the claim of his 2009 return. In contrast, Chrono Fact organizes the evidence chronologically, recognizing Georgi Andonov s return to Botev Plovdiv in 2009 after a hiatus, yielding a correct prediction. Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence (IJCAI-25) Claim: Yossi Yona began studying for a Ph D and went on to become a Professor of philosophy of education at Ben - Gurion University before he joined the Left Camp of Israel party . Label: REF Claim Events c1: Yossi Yona began studying for a Ph D Claim Chrono. Final c2: Yossi Yona become a Professor of philosophy of education at Ben - Gurion University Event Order Claim c3: Yossi Yona joined the Left Camp of Israel party Label TACV Evidence Events: After graduating in 1979, he began studying for a Ph D, graduating from the University of Pennsylvania in Philadelphia SUP c1 N.A. SUP Yossi Yona became a Professor of philosophy of education at Ben-Gurion University He subsequently joined the Education Department at Ben-Gurion University of the Negev. Whilst at university he joined the Left Camp of Israel party. SUP c3 Chrono Fact Evidence Events in Chronological Order: 1. Whilst at university, Yossi Yona joined the Left Camp of Israel party SUP c3 REF REF 2. After graduating in 1979, Yossi Yona began studying for a Ph D, graduating from the University of Pennsylvania in Philadelphia 3. He subsequently joined the Education Department at Ben-Gurion University of the Negev. 4. Yossi Yona became a Professor of philosophy of education at Ben-Gurion University SUP c2 Table 6: Sample T-FEVEROUS claim where chronological order of claim events is inconsistent with that of evidence events. Claim: Davie Dodds was a member of Dundee United F.C. before joining Arbroath F.C., thereafter becoming part of the Scotland national football team, then moving to Neuchˆatel Xamax, and finally playing for Aberdeen F.C. Label: SUP Claim Events c1: Davie Dodds was a member of Dundee United F.C. Claim Chrono. Final c2: Davie Dodds joined Arbroath F.C. after joining Dundee United F.C. Event Order Claim c3: Davie Dodds became part of the Scotland national football team Label c4: Davie Dodds moved to Neuchˆatel Xamax c5: Davie Dodds played for Aberdeen F.C. TACV Evidence Events: Davie Dodds is a member of the Dundee United F.C. from 1975 until 1986 SUP c1 N.A. REF Davie Dodds is a member of the Arbroath F.C. from 1977 until 1978 REF c2 Davie Dodds is a member of the Scotland national football from 1983 until 1983. SUP c3 Davie Dodds is a member of the Neuchˆatel Xamax from 1986 until 1986 SUP c4 Davie Dodds is a member of the Aberdeen F.C. from 1986 until 1989 SUP c5 Chrono Fact Evidence Events in Chronological Order: 1. Davie Dodds is a member of the Dundee United F.C. from 1975 until 1986 SUP c1 SUP SUP 2. Davie Dodds is a member of the Arbroath F.C. from 1977 until 1978 SUP c2 3. Davie Dodds is a member of the Scotland national football from 1983 until 1983 SUP c3 4. Davie Dodds is a member of the Neuchˆatel Xamax from 1986 until 1986 SUP c4 5. Davie Dodds is a member of the Aberdeen F.C. from 1986 until 1989 SUP c5 Table 7: Sample claim in Chrono Claims dataset involving overlapping events. Claim: Georgi Andonov was a member of the Botev Plovdiv from 2002 to 2006, joined the Bulgaria national under-21 team from 2003 to 2005, returned to Botev Plovdiv in 2009, played for PSFC Chernomorets Burgas from 2010 to 2012, and then became a member of PFC Beroe Stara Zagora starting in 2015. Label: SUP Claim Events c1: Georgi Andonov was a member of the Botev Plovdiv from 2002 to 2006 Claim Chrono. Final c2: Georgi Andonov joined the Bulgaria national under-21 team from 2003 to 2005. Event Order Claim c3: Georgi Andonov returned to Botev Plovdiv in 2009 Label c4: Georgi Andonov played for PSFC Chernomorets Burgas from 2010 to 2012 c5: Georgi Andonov became a member of PFC Beroe Stara Zagora starting in 2015 TACV Evidence Events: Georgi Andonov is a member of the Botev Plovdiv from 2002 until 2006 SUP c1 N.A. REF REF c3 Georgi Andonov is a member of the Bulgaria national under-21 team from 2003 until 2005 SUP c2 Georgi Andonov is a member of the PSFC Chernomorets Burgas from 2010 until 2012 SUP c4 Georgi Andonov is a member of the PFC Beroe Stara Zagora from 2015 SUP c5 Georgi Andonov is member of Botev Plovdiv from 2009 until 2009 Chrono Fact Evidence Events in Chronological Order: 1. Georgi Andonov is a member of the Botev Plovdiv from 2002 until 2006 SUP c1 SUP SUP 2. Georgi Andonov is a member of the Bulgaria national under-21 team from 2003 until 2005 SUP c2 3. Georgi Andonov is a member of the Botev Plovdiv from 2009 until 2009 SUP c3 4. Georgi Andonov is a member of the PSFC Chernomorets Burgas from 2010 until 2012 SUP c4 5. Georgi Andonov is a member of the PFC Beroe Stara Zagora from 2015 SUP c5 Table 8: Sample claim in Chrono Claims dataset involving recurring event. 7 Conclusion We have introduced a framework for temporal claim verification that incorporates temporal reasoning based on the chronological order of events. By utilizing a multi-level attention encoder, Chrono Fact effectively captures the relevance of evidence events to claim events. This enables Chrono Fact to accurately predict to claim event labels and verify the chronological order consistency between claim and evidence events before determining the overall claim label. We have curated a new benchmark dataset that involves complex claims with multiple events that may overlap or recur. Extensive experiments on multiple datasets have shown that Chrono Fact outperforms state-of-the-art models. Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence (IJCAI-25) Acknowledgments This work is supported by the Ministry of Education, Singapore, under its MOE Ac RF TIER 3 Grant (MOEMOET32022-0001). References [Allein et al., 2021] Liesbeth Allein, Isabelle Augenstein, and Marie-Francine Moens. Time-aware evidence ranking for fact-checking. Journal of Web Semantics, 71:100663, 2021. [Allein et al., 2023] Liesbeth Allein, Marlon Saelens, Ruben Cartuyvels, and Marie Francine Moens. Implicit temporal reasoning for evidence-based fact-checking. In Findings of the Association for Computational Linguistics: EACL 2023, pages 176 189, 2023. [Allen, 1983] James F Allen. Maintaining knowledge about temporal intervals. Communications of the ACM, 26(11):832 843, 1983. [Aly et al., 2021] Rami Aly, Zhijiang Guo, Michael Sejr Schlichtkrull, James Thorne, Andreas Vlachos, Christos Christodoulopoulos, Oana Cocarascu, and Arpit Mittal. FEVEROUS: Fact extraction and VERification over unstructured and structured information. 2021. [Barik et al., 2022] Anab Maulana Barik, Wynne Hsu, and Mong Li Lee. Incorporating external knowledge for evidence-based fact verification. In Companion Proceedings of the Web Conference 2022, pages 429 437, 2022. [Barik et al., 2024] Anab Barik, Wynne Hsu, and Mong-Li Lee. Time matters: An end-to-end solution for temporal claim verification. In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing: Industry Track, pages 657 664, 2024. [Chung et al., 2024] Hyung Won Chung, Le Hou, Shayne Longpre, Barret Zoph, Yi Tay, William Fedus, Yunxuan Li, Xuezhi Wang, Mostafa Dehghani, Siddhartha Brahma, et al. Scaling instruction-finetuned language models. Journal of Machine Learning Research, 25(70):1 53, 2024. [De Cao et al., 2020] N De Cao, G Izacard, S Riedel, and F Petroni. Autoregressive entity retrieval. In ICLR 20219th International Conference on Learning Representations, volume 2021. ICLR, 2020. [Jia et al., 2024] Zhen Jia, Philipp Christmann, and Gerhard Weikum. Faithful temporal question answering over heterogeneous sources. In Proceedings of the ACM on Web Conference 2024, pages 2052 2063, 2024. [Liu et al., 2020] Zhenghao Liu, Chenyan Xiong, Maosong Sun, and Zhiyuan Liu. Fine-grained fact verification with kernel graph attention network. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 7342 7351, 2020. [Mori et al., 2022] Marco Mori, Paolo Papotti, Luigi Bellomarini, and Oliver Giudice. Neural machine translation for fact-checking temporal claims. In Proceedings of the Fifth Fact Extraction and VERification Workshop (FEVER), pages 78 82, 2022. [Pan et al., 2023] Liangming Pan, Xinyuan Lu, Min-Yen Kan, and Preslav Nakov. Qacheck: A demonstration system for question-guided multi-hop fact-checking. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing: System Demonstrations, pages 264 273, 2023. [Qudus et al., 2023] Umair Qudus, Michael R oder, Sabrina Kirrane, and Axel-Cyrille Ngonga Ngomo. Temporalfc: A temporal fact checking approach over knowledge graphs. In International Semantic Web Conference, pages 465 483. Springer, 2023. [Soleimani et al., 2020] A Soleimani, C Monz, and M Worring. Bert for evidence retrieval and claim verification. Advances in Information Retrieval, 12036:359 366, 2020. [Stammbach and Neumann, 2019] Dominik Stammbach and Guenter Neumann. Team domlin: Exploiting evidence enhancement for the fever shared task. In Proceedings of the Second Workshop on Fact Extraction and VERification (FEVER), pages 105 109, 2019. [Thorne et al., 2018a] James Thorne, Andreas Vlachos, Christos Christodoulopoulos, and Arpit Mittal. Fever: a large-scale dataset for fact extraction and verification. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers), pages 809 819, 2018. [Thorne et al., 2018b] James Thorne, Andreas Vlachos, Oana Cocarascu, Christos Christodoulopoulos, and Arpit Mittal. The FEVER2.0 shared task. In Proceedings of the Second Workshop on Fact Extraction and VERification (FEVER), 2018. [Vaswani et al., 2017] Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. Attention is all you need. Advances in neural information processing systems, 30, 2017. [Venktesh et al., 2024] V Venktesh, Abhijit Anand, Avishek Anand, and Vinay Setty. Quantemp: A real-world opendomain benchmark for fact-checking numerical claims. In 47th International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR 2024, pages 650 660. Association for Computing Machinery (ACM), 2024. [Vrandeˇci c and Kr otzsch, 2014] Denny Vrandeˇci c and Markus Kr otzsch. Wikidata: a free collaborative knowledgebase. Communications of the ACM, 57(10):78 85, 2014. [Zhong et al., 2020] Wanjun Zhong, Jingjing Xu, Duyu Tang, Zenan Xu, Nan Duan, Ming Zhou, Jiahai Wang, and Jian Yin. Reasoning over semantic-level graph for fact checking. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 6170 6180, 2020. [Zhou et al., 2019] Jie Zhou, Xu Han, Cheng Yang, Zhiyuan Liu, Lifeng Wang, Changcheng Li, and Maosong Sun. Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence (IJCAI-25) Gear: Graph-based evidence aggregating and reasoning for fact verification. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 892 901, 2019. Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence (IJCAI-25)