Large-Scale Pretraining for Visual Dialog: A Simple State-of-the-Art Baseline

Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 12363)


Prior work in visual dialog has focused on training deep neural models on VisDial in isolation. Instead, we present an approach to leverage pretraining on related vision-language datasets before transferring to visual dialog. We adapt the recently proposed ViLBERT model for multi-turn visually-grounded conversations. Our model is pretrained on the Conceptual Captions and Visual Question Answering datasets, and finetuned on VisDial. Our best single model outperforms prior published work by \(1\%\) absolute on NDCG and MRR.

Next, we find that additional finetuning using “dense” annotations in VisDial leads to even higher NDCG – more than \(10\%\) over our base model – but hurts MRR – more than \(17\%\) below our base model! This highlights a trade-off between the two primary metrics – NDCG and MRR – which we find is due to dense annotations not correlating well with the original ground-truth answers to questions.


Vision & Language Visual dialog 



The Georgia Tech effort was supported in part by NSF, AFRL, DARPA, ONR YIPs, ARO PECASE, Amazon. AD was supported by fellowships from Facebook, Adobe, Snap Inc. Views and conclusions contained herein are those of the authors and should not be interpreted as necessarily representing the official policies or endorsements, either expressed or implied, of the U.S. Government, or any sponsor.

Supplementary material

504473_1_En_20_MOESM1_ESM.pdf (1.7 mb)
Supplementary material 1 (pdf 1763 KB)


  1. 1.
    de Vries, H., Strub, F., Chandar, S., Pietquin, O., Larochelle, H., Courville, A.: GuessWhat?! visual object discovery through multi-modal dialogue. In: CVPR (2017)Google Scholar
  2. 2.
    Das, A., et al.: Visual dialog. In: CVPR (2017)Google Scholar
  3. 3.
    Strub, F., De Vries, H., Mary, J., Piot, B., Courville, A., Pietquin, O.: End-to-end optimization of goal-driven and visually grounded dialogue systems, arXiv preprint arXiv:1703.05423 (2017)
  4. 4.
    Das, A., Kottur, S., Moura, J.M., Lee, S., Batra, D.: Learning cooperative visual dialog agents with deep reinforcement learning. In: ICCV (2017)Google Scholar
  5. 5.
    Lu, J., Kannan, A., Yang, J., Parikh, D., Batra, D.: Best of both worlds: transferring knowledge from discriminative learning to a generative visual dialog model. In: NIPS (2017)Google Scholar
  6. 6.
    Massiceti, D., Siddharth, N., Dokania, P.K., Torr, P.H.: FLIPDIAL: a generative model for two-way visual dialogue. In: CVPR (2018)Google Scholar
  7. 7.
    Wu, Q., Wang, P., Shen, C., Reid, I., van den Hengel, A.: Are you talking to me? reasoned visual dialog generation through adversarial learning. In: CVPR (2018)Google Scholar
  8. 8.
    Jain, U., Lazebnik, S., Schwing, A.G.: Two can play this game: visual dialog with discriminative question generation and answering. In: CVPR (2018)Google Scholar
  9. 9.
    Kottur, S., Moura, J.M., Parikh, D., Batra, D., Rohrbach, M.: Visual coreference resolution in visual dialog using neural module networks. In: ECCV (2018)Google Scholar
  10. 10.
    Lee, S.-W., Gao, T., Yang, S., Yoo, J., Ha, J.-W.: Large-scale answerer in questioner’s mind for visual dialog question generation. In: ICLR (2019)Google Scholar
  11. 11.
    Niu, Y., Zhang, H., Zhang, M., Zhang, J., Lu, Z., Wen, J.-R.: Recursive visual attention in visual dialog. In: CVPR (2019)Google Scholar
  12. 12.
    Zheng, Z., Wang, W., Qi, S., Zhu, S.-C.: Reasoning visual dialogs with structural and partial observations. In: CVPR (2019)Google Scholar
  13. 13.
    Schwartz, I., Yu, S., Hazan, T., Schwing, A.G.: Factor graph attention. In: CVPR (2019)Google Scholar
  14. 14.
    Kang, G.-C., Lim, J., Zhang, B.-T.: Dual attention networks for visual reference resolution in visual dialog. In: EMNLP (2019)Google Scholar
  15. 15.
    Gan, Z., Cheng, Y., Kholy, A.E., Li, L., Liu, J., Gao, J.: Multi-step reasoning via recurrent dual attention for visual dialog. In: ACL (2019)Google Scholar
  16. 16.
    Kottur, S., Moura, J.M., Parikh, D., Batra, D., Rohrbach, M.: CLEVR-dialog: a diagnostic dataset for multi-round reasoning in visual dialog. In: NAACL (2019)Google Scholar
  17. 17.
    Murahari, V., Chattopadhyay, P., Batra, D., Parikh, D., Das, A.: Improving generative visual dialog by answering diverse questions. In: EMNLP (2019)Google Scholar
  18. 18.
    Shekhar, R., et al.: Beyond task success: a closer look at jointly learning to see, ask, and guesswhat. In: NAACL (2019)Google Scholar
  19. 19.
    Yang, T., Zha, Z.-J., Zhang, H.: Making history matter: gold-critic sequence training for visual dialog. arXiv preprint arXiv:1902.09326 (2019)
  20. 20.
    Guo, D., Xu, C., Tao, D.: Image-question-answer synergistic network for visual dialog. In: CVPR (2019)Google Scholar
  21. 21.
    Qi, J., Niu, Y., Huang, J., Zhang, H.: Two causal principles for improving visual dialog, arXiv preprint arXiv:1911.10496 (2019)
  22. 22.
    Jiang, X., Yu, J., et al.: DualVD: an adaptive dual encoding model for deep visual understanding in visual dialogue. In: AAAI (2020)Google Scholar
  23. 23.
    Alamri, H., et al.: Audio visual scene-aware dialog. In: CVPR (2019)Google Scholar
  24. 24.
    de Vries, H., Shuster, K., Batra, D., Parikh, D., Weston, J., Kiela, D.: Talk the walk: navigating new york city through grounded dialogue, arXiv preprint arXiv:1807.03367 (2018)
  25. 25.
    Nguyen, K., Daumé III, H.: Help, anna! visual navigation with natural multimodal assistance via retrospective curiosity-encouraging imitation learning. In: EMNLP (2019)Google Scholar
  26. 26.
    Thomason, J., Murray, M., Cakmak, M., Zettlemoyer, L.: Vision-and-dialog navigation (2019)Google Scholar
  27. 27.
    Zhu, Y., Kiros, R., et al.: Aligning books and movies: towards story-like visual explanations by watching movies and reading books. In: ICCV (2015)Google Scholar
  28. 28.
    Sharma, P., Ding, N., Goodman, S., Soricut, R.: Conceptual captions: a cleaned, hypernymed, image alt-text dataset for automatic image captioning. In: ACL (2018)Google Scholar
  29. 29.
    Antol, S., et al.: VQA: visual question answering. In: ICCV (2015)Google Scholar
  30. 30.
    He, K., Zhang, X., Ren, S., Sun, J.: Deep residual learning for image recognition. In: CVPR (2016)Google Scholar
  31. 31.
    Simonyan, K., Zisserman, A.: Very deep convolutional networks for large-scale image recognition. In: ICLR (2015)Google Scholar
  32. 32.
    Krizhevsky, A., Sutskever, I., Hinton, G.: ImageNet classification with deep convolutional neural networks. In: NIPS (2012)Google Scholar
  33. 33.
    Radford, A., Narasimhan, K., Salimans, T., Sutskever, I.: Improving language understanding with unsupervised learning (2018)Google Scholar
  34. 34.
    Vaswani, A., et al.: Attention is all you need. In: NIPS (2017)Google Scholar
  35. 35.
    Devlin, J., Chang, M.-W., Lee, K., Toutanova, K.: BERT: pre-training of deep bidirectional transformers for language understanding. In: NAACL (2019)Google Scholar
  36. 36.
    Liu, Y., et al.: RoBERTa: a robustly optimized BERT pretraining approach, arXiv preprint arXiv:1907.11692 (2019)
  37. 37.
    Lan, Z., Chen, M., Goodman, S., Gimpel, K., Sharma, P., Soricut, R.: ALBERT: a lite BERT for self-supervised learning of language representations, arXiv preprint arXiv:1909.11942 (2019)
  38. 38.
    Yang, Z., Dai, Z., Yang, Y., Carbonell, J., Salakhutdinov, R., Le, Q.L.: XLNET: generalized autoregressive pretraining for language understanding, arXiv preprint arXiv:1906.08237 (2019)
  39. 39.
    Raffel, C., et al.: Exploring the limits of transfer learning with a unified text-to-text transformer, arXiv preprint arXiv:1910.10683 (2019)
  40. 40.
    Zhang, Y., et al.: DialoGPT: large-scale generative pre-training for conversational response generation, arXiv preprint arXiv:1911.00536 (2019)
  41. 41.
    Russakovsky, O., et al.: ImageNet large scale visual recognition challenge. IJCV (2015)Google Scholar
  42. 42.
    Krishna, R., et al.: Visual genome: connecting language and vision using crowdsourced dense image annotations. IJCV (2017)Google Scholar
  43. 43.
    Wang, A., Singh, A., Michael, J., Hill, F., Levy, O., Bowman, S.R.: GLUE: a multi-task benchmark and analysis platform for natural language understanding, arXiv preprint arXiv:1804.07461 (2018)
  44. 44.
    Lu, J., Batra, D., Parikh, D., Lee, S.: ViLBERT: pretraining task-agnostic visiolinguistic representations for vision-and-language tasks. In: NeurIPS (2019)Google Scholar
  45. 45.
    Li, L.H., Yatskar, M., Yin, D., Hsieh, C.-J., Chang, K.-W.: VisualBERT: a simple and performant baseline for vision and language, arXiv preprint arXiv:1908.03557 (2019)
  46. 46.
    Tan, H., Bansal, M.: LXMERT: learning cross-modality encoder representations from transformers, arXiv preprint arXiv:1908.07490 (2019)
  47. 47.
    Chen, Y.-C., et al.: UNITER: Learning UNiversal Image-TExt Representations, arXiv preprint arXiv:1909.11740 (2019)
  48. 48.
    Li, G., Duan, N., Fang, Y., Jiang, D., Zhou, M.: Unicoder-VL: a universal encoder for vision and language by cross-modal pre-training, arXiv preprint arXiv:1908.06066 (2019)
  49. 49.
    Su, W., et al.: VL-BERT: pre-training of generic visual-linguistic representations, arXiv preprint arXiv:1908.08530 (2019)
  50. 50.
    Sun, C., Myers, A., Vondrick, C., Murphy, K., Schmid, C.: VideoBERT: a joint model for video and language representation learning, arXiv preprint arXiv:1904.01766 (2019)
  51. 51.
    Zellers, R., Bisk, Y., Farhadi, A., Choi, Y.: From recognition to cognition: visual commonsense reasoning. In: CVPR (2019)Google Scholar
  52. 52.
    Suhr, A., Zhou, S., Zhang, A., Zhang, I., Bai, H., Artzi, Y.: A corpus for reasoning about natural language grounded in photographs. In: ACL (2019)Google Scholar
  53. 53.
    Xie, N., Lai, F., Doran, D., Kadav, A.: Visual entailment: a novel task for fine-grained image understanding, arXiv preprint arXiv:1901.06706 (2019)
  54. 54.
    Young, P., Lai, A., Hodosh, M., Hockenmaier, J.: From image descriptions to visual denotations: new similarity metrics for semantic inference over event descriptions. In: TACL (2014)Google Scholar
  55. 55.
    Lee, K.-H., Chen, X., Hua, G., Hu, H., He, X.: Stacked cross attention for image-text matching. In: ECCV (2018)Google Scholar
  56. 56.
    Kazemzadeh, S., Ordonez, V., Matten, M., Berg, T.L.: ReferItGame: referring to objects in photographs of natural scenes. In: EMNLP (2014)Google Scholar
  57. 57.
    Hao, W., Li, C., Li, X., Carin, L., Gao, J.: Towards learning a generic agent for vision-and-language navigation via pre-training. In: CVPR (2020)Google Scholar
  58. 58.
    Paszke, A., et al.: Pytorch: an imperative style, high-performance deep learning library. In: Wallach, H., Larochelle, H., Beygelzimer, A., d’Alché-Buc, F., Fox, E., Garnett, R. (eds.) Advances in Neural Information Processing Systems, vol. 32, pp. 8024–8035, Curran Associates Inc. (2019)Google Scholar
  59. 59.
    Wolf, T., Sanh, V., Chaumond, J., Delangue, C.: TransferTransfo: a transfer learning approach for neural network based conversational agents, arXiv preprint arXiv:1901.08149 (2019)
  60. 60.
    Anderson, P., et al.: Bottom-up and top-down attention for image captioning and visual question answering, arXiv preprint arXiv:1707.07998 (2017)
  61. 61.
    Ren, S., He, K., Girshick, R., Sun, J.: Faster R-CNN: towards real-time object detection with region proposal networks. In: NIPS (2015)Google Scholar
  62. 62.
    Lin, T.-Y., et al.: Microsoft COCO: common objects in context. In: ECCV (2014)Google Scholar
  63. 63.
    Kingma, D., Ba, J.: Adam: a method for stochastic optimization. In: ICLR (2015)Google Scholar
  64. 64.
    Lewis, M., et al.: Bart: denoising sequence-to-sequence pre-training for natural language generation, translation, and comprehension, arXiv preprint arXiv:1910.13461 (2019)

Copyright information

© Springer Nature Switzerland AG 2020

Authors and Affiliations

  1. 1.Georgia Institute of TechnologyAtlantaUSA
  2. 2.Facebook AI ResearchMenlo ParkUSA

Personalised recommendations