Question matching is the fundamental task in retrieval-based dialogue systems which assesses the similarity between Query and Question. Unfortunately, existing methods focus on improving the accuracy of text similarity in the general domain, without adaptation to the financial domain. Financial question matching has two critical issues: (1) How to accurately model the contextual representation of a financial sentence? (2) How to accurately represent financial key phrases in an utterance? To address these issues, this paper proposes a novel Financial Knowledge Enhanced Network (FinKENet) that significantly injects financial knowledge into contextual text. Specifically, we propose a multi-level encoder to extract both sentence-level features and financial phrase-level features, which can more accurately represent sentences and financial phrases. Furthermore, we propose a financial co-attention adapter to combine sentence features and financial keyword features. Finally, we design a multi-level similarity decoder to calculate the similarity between queries and questions. In addition, a cross-entropy-based loss function is presented for model optimization. Experimental results demonstrate the effectiveness of the proposed method on the Ant Financial question matching dataset. In particular, the Recall score improves from 73.21% to 74.90% (1.69% absolute).

Binaural audio is essential for delivering immersive spatial auditory experiences through headsets. However, due to the high cost and complexity of binaural recording, there has been growing research interest in binaural audio synthesis (BAS) from monaural inputs. In natural listening environments, humans typically perceive multiple concurrent sound sources, yet most existing BAS approaches render each source independently, relying on perfect source signal separation, a condition rarely achievable in practice and often leading to perceptual quality degradation. To address this limitation, this paper proposes MixBAS, a transformer based end-to-end multi-source mono-to-binaural synthesis framework that eliminates the need for explicit source separation. We design an asymmetric transformer that spatializes a mono mixture, which comprises both speech and non-speech components, into its binaural counterpart by incorporating a user-defined positional prompt for the non-speech source. When reproduced over headphones, the generated binaural audio enables listeners to perceive a high-quality speech signal along with a non-speech source rendered at a user-specified spatial location. Experimental results demonstrate that MixBAS significantly outperforms existing BAS baselines relying on source separation in both objective metrics and perceptual quality.