In-context learning is governed by both temporal and semantic relationships, shaping how Large Language Models (LLMs) retrieve contextual information. Analogous to human episodic memory, where the retrieval of specific events is enabled by separating events that happened at different times, this work probes the ability of various pretrained LLMs, including transformer and state-space models, to differentiate and retrieve temporally separated events. Specifically, we prompted models with sequences containing multiple presentations of the same token, which reappears at the sequence end. By fixing the positions of these repeated tokens and permuting all others, we removed semantic confounds and isolated temporal effects on next-token prediction. Across diverse sequences, models consistently placed the highest probabilities on tokens following a repeated token, but with a notable bias for those nearest the beginning or end of the input. An ablation experiment linked this phenomenon in transformers to induction heads. Extending the analysis to unique semantic contexts with partial overlap further demonstrated that memories embedded in the middle of a prompt are retrieved less reliably. Despite architectural differences, state-space and transformer models showed comparable temporal biases. Our findings deepen the understanding of temporal biases in in-context learning and offer an illustration of how these biases can enable temporal separation and episodic retrieval.

Large Audio Language Models (LALMs) are increasingly applied to audio understanding and multimodal reasoning, yet their ability to locate when events occur remains underexplored. We present the first systematic study of temporal bias in LALMs, revealing a key limitation in their timestamp prediction. For example, when asked "At which second does the lecturer introduce the key formula?", models often predict timestamps that are consistently earlier or later than the ground truth. Through controlled experiments on timestamped datasets, we find that temporal bias (i) is prevalent across datasets and models, (ii) increases with audio length - even accumulating to tens of seconds in extended recordings, and (iii) varies across event types and positions. We quantify this effect with the Temporal Bias Index (TBI), measuring systematic misalignment in predicted event timings, and complement it with a visualization framework. Our findings highlight a fundamental limitation in current LALMs and call for the development of temporally robust architectures.

This paper introduces DateLogicQA, a benchmark with 190 questions covering diverse date formats, temporal contexts, and reasoning types. We propose the Semantic Integrity Metric to assess tokenization quality and analyse two biases: Representation-Level Bias, affecting embeddings, and Logical-Level Bias, influencing reasoning outputs. Our findings provide a comprehensive evaluation of LLMs' capabilities and limitations in temporal reasoning, highlighting key challenges in handling temporal data accurately.

Efficient learning from streaming data is important for modern data analysis due to the continuous and rapid evolution of data streams. Despite significant advancements in stream pattern mining, challenges persist, particularly in managing complex data streams like sequential and weighted itemsets. While reservoir sampling serves as a fundamental method for randomly selecting fixed-size samples from data streams, its application to such complex patterns remains largely unexplored. In this study, we introduce an approach that harnesses a weighted reservoir to facilitate direct pattern sampling from streaming batch data, thus ensuring scalability and efficiency. We present a generic algorithm capable of addressing temporal biases and handling various pattern types, including sequential, weighted, and unweighted itemsets. Through comprehensive experiments conducted on real-world datasets, we evaluate the effectiveness of our method, showcasing its ability to construct accurate incremental online classifiers for sequential data. Our approach not only enables previously unusable online machine learning models for sequential data to achieve accuracy comparable to offline baselines but also represents significant progress in the development of incremental online sequential itemset classifiers.

Previous work identified temporal bias in an Italian hate In recent years, researchers have developed a huge variety speech data set associated with immigrants (Florio et al. of machine learning models that can automatically detect 2020). However, they have yet to explore temporal factors abusive language (Mishra et al. 2019; Aurpa, Sadik, and affecting predictive performance from a multilingual perspective. Ahmed 2022; Das and Mukherjee 2023; Alrashidi, Jamal, In this paper, we explore temporal bias in 5 different and Alkhathlan 2023). However, these models may be subject abusive data sets that span varying time periods, in 4 to temporal bias, which can lead to a decrease in the languages (English, Spanish, Italian, and Chinese). Specifically, accuracy of abusive language detection models, potentially we investigate the following core research questions: allowing abusive language to be undetected or falsely detected. RQ1: How does the magnitude of temporal bias vary across different data sets such as language, time span and Temporal bias arises from differences in populations and collection methods?

Written by Camylle Lanteigne (@CamLante), who's currently pursuing a Master's in Public Policy at Concordia University and whose work on social robots and empathy has been featured on Vox. This explainer was written in response to colleagues' requests to know more about temporal bias in AI ethics. It begins with a refresher on cognitive biases, then dives into: how humans understand time, time preferences, present-day preference, confidence changes, planning fallacies, and hindsight bias. Bias is a really big topic, but I'll try to succinctly define a subsection of it--implicit cognitive bias--in a way that is useful for AI ethics, particularly. Humans have cognitive biases, which means every one of us, to varying degrees, holds beliefs and impressions that are not backed up by fleshed out reasoning or evidence, or that we never bothered questioning in the first place.¹