Papers by Xianchao Wu
Dialog Generation Using Multi-Turn Reasoning Neural Networks (N18-1)
Copied to clipboard
| Challenge: | Existing methods for dialog generation are limited and short at generalization. |
| Approach: | They propose a generalizable dialog generation approach that adapts multi-turn reasoning to generate responses by taking current conversation session context as a document and current query as 'question' they separate the single memory used for document comprehension into different groups for speaker-specific topic and opinion embedding. |
| Outcome: | Experiments on Japanese 10-sentence (5-round) conversation modeling show that multi-turn reasoning can produce more diverse and acceptable responses than state-of-the-art single-turn and non-reasoning baselines. |
Playing 20 Question Game with Policy-Based Reinforcement Learning (D18-1)
Copied to clipboard
| Challenge: | The 20 Questions (Q20) game encourages deductive reasoning and creativity. |
| Approach: | They propose a policy-based Reinforcement Learning method which learns optimal question selection . the method is robust to noisy answers and uses a reward network to estimate the more informative reward . |
| Outcome: | The proposed method outperforms an entropy-based engineering system and has competitive performance in noisy-free simulation environment. |
Duplex Diffusion Models Improve Speech-to-Speech Translation (2023.findings-acl)
Copied to clipboard
| Challenge: | Existing approaches to speech-to-speech translation train two separate models or a multitask-learned model with low efficiency and inferior performance. |
| Approach: | They propose a duplex diffusion model that applies diffusion probabilistic models to both sides of a reversible duplex Conformer and enables reverse speech translation by simply flipping the input and output ends. |
| Outcome: | The proposed model achieves the first success of reversible speech translation with significant improvements of ASR-BLEU scores compared with a list of state-of-the-art baselines. |
Event-Driven Learning of Systematic Behaviours in Stock Markets (2020.findings-emnlp)
Copied to clipboard
| Challenge: | Using financial news, we can predict stock market behaviours by extracting financial events from the news and ranking the importance of the events. |
| Approach: | They propose to combine open information extraction and neural co-reference resolution to extract financial events from news streams and extend hierarchical attention networks that include attentions on event, news and temporal levels. |
| Outcome: | The proposed method achieves significantly better accuracies and higher simulated annualized returns than state-of-the-art models when being applied to predicting Standard&Poor 500, Dow Jones, Nasdaq indices and 10 individual stocks. |
SteerLM: Attribute Conditioned SFT as an (User-Steerable) Alternative to RLHF (2023.findings-emnlp)
Copied to clipboard
| Challenge: | supervised fine-tuning and reinforcement learning from human feedback (RLHF) are not effective in generating useful and high-quality responses. |
| Approach: | They propose a supervised fine-tuning method that empowers end-users to control responses during inference. |
| Outcome: | Experiments show that supervised fine-tuning and reinforcement learning from human feedback (RLHF) can generate helpful and high-quality responses while maintaining customizability. |