Learning to Embed Multi-Modal Contexts for Situated Conversational Agents
The Situated Interactive Multi-Modal Conversations (SIMMC) 2.0 aims to create virtual shopping assistants that can accept complex multi-modal inputs, i.e. visual appearances of objects and user utterances. It consists of four subtasks, multi-modal disambiguation (MM-Disamb), multi-modal coreference resolution (MM-Coref), multi-modal dialog state tracking (MM-DST), and response retrieval and generation. While many task-oriented dialog systems usually tackle each subtask separately, we propose a jointly learned multi-modal encoder-decoder that incorporates visual inputs and performs all four subtasks at once for efficiency. This approach won the MM-Coref and response retrieval subtasks and nominated runner-up for the remaining subtasks using a single unified model at the 10th Dialog Systems Technology Challenge (DSTC10), setting a high bar for the novel task of multi-modal task-oriented dialog systems.
PDF Abstract
SIMMC
SIMMC2.0
| Task | Dataset | Model | Metric Name | Metric Value | Global Rank | Benchmark |
|---|---|---|---|---|---|---|
| Response Generation | SIMMC2.0 | BART-large | BLEU | 33.1 | # 2 | |
| Dialogue State Tracking | SIMMC2.0 | BART-base | Slot F1 | 82.0 | # 3 | |
| Act F1 | 95.2 | # 3 | ||||
| Dialogue State Tracking | SIMMC2.0 | BART-large | Slot F1 | 88.3 | # 1 | |
| Act F1 | 96.3 | # 2 |
