The modern digital world and associated innovative and state-of-the-art applications that characterize its presence, render the current digital age a captivating era for many worldwide. These innovations include dialogue systems, such as Apple’s Siri, Google Now, and Microsoft’s Cortana, that stay on the personal devices of users and assist them in their daily activities. These systems track the intentions of users by analyzing their speech, context by looking at their previous turns, and several other external details, and respond or act in the form of speech output. For these systems to work efficiently, a dialogue state tracking (DST) module is required to infer the current state of the dialogue in a conversation by processing previous states up to the current state. However, developing a DST module that tracks and exploit dialogue states effectively and accurately is challenging. The notable challenges that warrant immediate attention include scalability, handling the unseen slot-value pairs during training, and retraining the model with changes in the domain ontology. In this article, we present a new end-to-end framework by combining BERT, Stacked Bidirectional LSTM (BiLSTM), and a multiple attention mechanism to formalize DST as a classification problem and address the aforementioned issues. The BERT-based module encodes the user’s and system’s utterances. The Stacked BiLSTM extracts the contextual features and multiple attention mechanisms to calculate the attention between its hidden states and the utterance embeddings. We experimentally evaluated our method against the current approaches over a variety of datasets. The results indicate a significant overall improvement. The proposed model is scalable in terms of sharing the parameters and it considers the unseen instances during training.

Coordinating lifelong collaboration among multiple agents poses significant challenges, particularly when agents must work together to complete tasks in dynamic environments. Existing methods typically separate task allocation from path planning, resulting in inefficient coordination, idle agents, and frequent deadlocks when collaboration is required. These issues are particularly critical in real-world domains, such as warehouse logistics, where multi-agent cooperation is crucial for handling large or complex tasks. To overcome these limitations, we propose the Lifelong Multi-Agent Collaborative Path Finding (LMACPF) as a unified framework, which tightly integrates task assignment and path planning for lifelong, collaboration-intensive scenarios. LMACPF combines Multi-Level Task Assignment (MLTA) with Collaborative Agent Search (CAS) and Collaborative Revised Prioritized Planning (CRPP) to optimize task allocation and path planning, particularly for cooperative agent behaviors. MLTA hierarchically assigns tasks based on collaboration requirements to ensure that multi-agent tasks are prioritized. CAS and CRPP enable synchronized planning and formation maintenance among agents. In addition, the Evaluation-Based Collaborative Large Neighborhood Search (EBC-LNS) is presented for a task-level evaluation that prioritizes collaborative agent groups and balances the path quality with real-time responsiveness. Experiments show MLTA increases task-ready agents by 45.8% over MAPD. LMACPF improves throughput and task success over state-of-the-art methods, demonstrating its effectiveness in scalable, lifelong multi-agent collaboration.