Conversational agents often encounter ambiguous user requests, requiring an effective clarification to successfully complete tasks. While recent advancements in real-world applications favor multi-agent architectures to manage complex conversational scenarios efficiently, ambiguity resolution remains a critical and underexplored challenge—particularly due to the difficulty of determining which agent should initiate a clarification and how agents should coordinate their actions when faced with uncertain or incomplete user input. The fundamental questions of when to interrupt a user and how to formulate the optimal clarification query within the most optimal multi-agent settings remain open. In this paper, we propose MAC (Multi-Agent Clarification), an interactive multi-agent framework specifically optimized to resolve user ambiguities by strategically managing clarification dialogues. We first introduce a novel taxonomy categorizing user ambiguities to systematically guide clarification strategies. Then, we present MAC that autonomously coordinates multiple agents to interact synergistically with users. Empirical evaluations on MultiWOZ 2.4 demonstrate that enabling clarification at both levels increases task success rate 7.8% (54.5 → 62.3) and reduces the average number of dialogue turns (6.53 → 4.86) by eliciting all required user information up front and minimizing repetition. Our findings highlight the importance of active user interaction and role-aware clarification for more reliable human–agent communication.

Effective human-agent collaboration is increasingly prevalent in real-world applications. Current trends in such collaborations are predominantly unidirectional, with users providing instructions or posing questions to agents, where agents respond directly without seeking necessary clarifications or confirmations. However, the evolving capabilities of these agents require more proactive engagement, where agents should dynamically participate in conversations to clarify user intents, resolve ambiguities, and adapt to changing circumstances. Existing prior work under-utilize the conversational capabilities of language models (LMs), thereby optimizing agents as better followers rather than effective speakers. In this work, we introduce SpeakRL, a reinforcement learning (RL) method that enhances agents’ conversational capabilities by rewarding proactive interactions with users, such as asking right clarification questions when necessary. To support this, we curate SpeakER, a synthetic dataset that includes diverse scenarios from task-oriented dialogues, where tasks are resolved through interactive clarification questions. We present a systematic analysis of reward design for conversational proactivity and propose a principled reward formulation for teaching agents to balance asking with acting. Empirical evaluations demonstrate that our approach achieves a 20.14% absolute improvement in task completion over base models without increasing conversation turns even surpassing even much larger proprietary models, demonstrating the promise of clarification-centric user-agent interactions.

In recent months, substantial progress has been made in complex reasoning of Large Language Models (LLMs), particularly through the application of test-time scaling. Notable examples include, though are not limited to, OpenAI’s o1/o3/o4 series and DeepSeek-R1. When responding to a query, these models generate an extended reasoning trajectory, during which the model explores, reflects, backtracks, and self-verifies before arriving at a conclusion. However, fine-tuning models with such reasoning trajectories may not always be optimal. Our findings indicate that not all components within these reasoning trajectories contribute positively to the reasoning process; in fact, some components may affect the overall performance negatively. In this study, we divide a reasoning trajectory into individual subtrajectories and develop a “5+2” framework to: (1) systematically identify suboptimal subtrajectories within the reasoning trajectory based on five human-established criteria; (2) assess the independence of the suboptimal subtrajectories identified in (1) from the subsequent content, ensuring that their elimination does not compromise overall flow and coherence of the reasoning process. Additionally, a sampling algorithm, built upon the “5+2” framework, is employed to select data whose reasoning process is free from suboptimal subtrajectories to the highest degree. Experimental results demonstrate that our method can reduce the number of suboptimal subtrajectories by 25.9% during the inference. Furthermore, our method achieves an average accuracy of 58.92% on highly challenging AIME24, AIME25, AMC24 and MATH500 benchmarks with only two thirds of training data, surpassing the average accuracy of 58.06% achieved with the entire data, and outperforming open-source datasets, including s1K-1.1, Light-R1-SFT-stage-1, OpenR1-Math-94k, and OpenThoughts-114k, when fine-tuning Qwen2.5-Math-7B. Finally, we have validated the efficacy of our method under resource-constrained scenarios, where it exhibits performance improvements across different maximum inference token limits: 2k, 4k, 8k, and 16k tokens.

The growing popularity of conversational AI agents such as Alexa, Google Assistant, and Siri rely on accurate spoken language comprehension. The query reformulation (QR) method, which reformulates defective user queries, has been broadly adopted to mitigate the challenges posed by understanding user’s intent from imperfect spoken recognition result. However, due to the scarcity of non-English QR labels, providing high-quality QR for non-English users still remains a challenge. This work proposes a novel cross-lingual QR framework, CL-QR, to leverage the abundant reformulation resources in English to improve non-English QR performance. The proposed work also proposes a Module-wise Mutually-supervised Feedback learning (MMF) algorithm to enable the continually self-improving of the CL-QR, which alleviates the lack of cross-lingual QR training data and enhances the delivery of high-quality reformulations learned in English for multilingual queries. Both offline evaluation and online A/B testing demonstrates the effectiveness of the proposed method.

Contextual query rewriting (CQR) is a crucial component in Conversational AI agents, leveraging the contextual information from previous user-agent conversations to improve the comprehension of current user intent. However, traditional CQR methods often concentrate on supervised fine-tuning only, neglecting the opportunities to learn from user feedback to align with user preferences. Inspired by recent advances in learning from human feedback (LHF), this paper proposes a novel Preference Aligned Contextual Query Rewriting (PA-CQR) framework to enhance the CQR model’s capability in generating user preference-aligned rewrites. This paper also investigates the efficacy of various state-of-the-art feedback learning algorithms on the CQR task, and proposes a novel Dynamic Direct Preference Optimization (Dynamic DPO) algorithm to better adapt the DPO algorithm to large-scale CQR training. Experiments on large-scale real-world CQR data set demonstrate the superiority of the proposed PA-CQR framework and the Dynamic DPO.

In recent years, Pre-trained Language Models (PLMs) have shown their superiority by pre-training on unstructured text corpus and then fine-tuning on downstream tasks. On entity-rich textual resources like Wikipedia, Knowledge-Enhanced PLMs (KEPLMs) incorporate the interactions between tokens and mentioned entities in pre-training, and are thus more effective on entity-centric tasks such as entity linking and relation classification. Although exploiting Wikipedia’s rich structures to some extent, conventional KEPLMs still neglect a unique layout of the corpus where each Wikipedia page is around a topic entity (identified by the page URL and shown in the page title). In this paper, we demonstrate that KEPLMs without incorporating the topic entities will lead to insufficient entity interaction and biased (relation) word semantics. We thus propose KEPLET, a novel Knowledge-Énhanced Pre-trained LanguagE model with Topic entity awareness. In an end-to-end manner, KEPLET identifies where to add the topic entity’s information in a Wikipedia sentence, fuses such information into token and mentioned entities representations, and supervises the network learning, through which it takes topic entities back into consideration. Experiments demonstrated the generality and superiority of KEPLET which was applied to two representative KEPLMs, achieving significant improvements on four entity-centric tasks.

Encoding both language-specific and language-agnostic information into a single high-dimensional space is a common practice of pre-trained Multi-lingual Language Models (pMLM). Such encoding has been shown to perform effectively on natural language tasks requiring semantics of the whole sentence (e.g., translation). However, its effectiveness appears to be limited on tasks requiring partial information of the utterance (e.g., multi-lingual entity retrieval, template retrieval, and semantic alignment). In this work, a novel Fine-grained Multilingual Disentangled Autoencoder (FMDA) is proposed to disentangle fine-grained semantic information from language-specific information in a multi-lingual setting. FMDA is capable of successfully extracting the disentangled template semantic and residual semantic representations. Experiments conducted on the MASSIVE dataset demonstrate that the disentangled encoding can boost each other during the training, thus consistently outperforming the original pMLM and the strong language disentanglement baseline on monolingual template retrieval and cross-lingual semantic retrieval tasks across multiple languages.

Self-learning paradigms in large-scale conversational AI agents tend to leverage user feedback in bridging between what they say and what they mean. However, such learning, particularly in Markov-based query rewriting systems have far from addressed the impact of these models on future training where successive feedback is inevitably contingent on the rewrite itself, especially in a continually updating environment. In this paper, we explore the consequences of this inherent lack of self-awareness towards impairing the model performance, ultimately resulting in both Type I and II errors over time. To that end, we propose augmenting the Markov Graph construction with a superposition-based adjacency matrix. Here, our method leverages an induced stochasticity to reactively learn a locally-adaptive decision boundary based on the performance of the individual rewrites in a bi-variate beta setting. We also surface a data augmentation strategy that leverages template-based generation in abridging complex conversation hierarchies of dialogs so as to simplify the learning process. All in all, we demonstrate that our self-aware model improves the overall PR-AUC by 27.45%, achieves a relative defect reduction of up to 31.22%, and is able to adapt quicker to changes in global preferences across a large number of customers.

Subword tokenization is a commonly used input pre-processing step in most recent NLP models. However, it limits the models’ ability to leverage end-to-end task learning. Its frequency-based vocabulary creation compromises tokenization in low-resource languages, leading models to produce suboptimal representations. Additionally, the dependency on a fixed vocabulary limits the subword models’ adaptability across languages and domains. In this work, we propose a vocabulary-free neural tokenizer by distilling segmentation information from heuristic-based subword tokenization. We pre-train our character-based tokenizer by processing unique words from multilingual corpus, thereby extensively increasing word diversity across languages. Unlike the predefined and fixed vocabularies in subword methods, our tokenizer allows end-to-end task learning, resulting in optimal task-specific tokenization. The experimental results show that replacing the subword tokenizer with our neural tokenizer consistently improves performance on multilingual (NLI) and code-switching (sentiment analysis) tasks, with larger gains in low-resource languages. Additionally, our neural tokenizer exhibits a robust performance on downstream tasks when adversarial noise is present (typos and misspelling), further increasing the initial improvements over statistical subword tokenizers.

Text Style Transfer (TST) aims to alter the underlying style of the source text to another specific style while keeping the same content. Due to the scarcity of high-quality parallel training data, unsupervised learning has become a trending direction for TST tasks. In this paper, we propose a novel VAE based Text Style Transfer with pivOt Words Enhancement leaRning (VT-STOWER) method which utilizes Variational AutoEncoder (VAE) and external style embeddings to learn semantics and style distribution jointly. Additionally, we introduce pivot words learning, which is applied to learn decisive words for a specific style and thereby further improve the overall performance of the style transfer. The proposed VT-STOWER can be scaled to different TST scenarios given very limited and non-parallel training data with a novel and flexible style strength control mechanism. Experiments demonstrate that the VT-STOWER outperforms the state-of-the-art on sentiment, formality, and code-switching TST tasks.