Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
102 tokens/sec
GPT-4o
59 tokens/sec
Gemini 2.5 Pro Pro
43 tokens/sec
o3 Pro
6 tokens/sec
GPT-4.1 Pro
50 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

History for Visual Dialog: Do we really need it? (2005.07493v1)

Published 8 May 2020 in cs.CV, cs.AI, cs.CL, and cs.LG

Abstract: Visual Dialog involves "understanding" the dialog history (what has been discussed previously) and the current question (what is asked), in addition to grounding information in the image, to generate the correct response. In this paper, we show that co-attention models which explicitly encode dialog history outperform models that don't, achieving state-of-the-art performance (72 % NDCG on val set). However, we also expose shortcomings of the crowd-sourcing dataset collection procedure by showing that history is indeed only required for a small amount of the data and that the current evaluation metric encourages generic replies. To that end, we propose a challenging subset (VisDialConv) of the VisDial val set and provide a benchmark of 63% NDCG.

User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (5)
  1. Shubham Agarwal (34 papers)
  2. Trung Bui (79 papers)
  3. Joon-Young Lee (61 papers)
  4. Ioannis Konstas (40 papers)
  5. Verena Rieser (58 papers)
Citations (67)

Summary

Overview of ACL 2020 Proceedings Instructions

The presented paper serves as a comprehensive guide for authors preparing their manuscripts for submission to the ACL 2020 conference. This document meticulously outlines the formatting and content expectations for both reviews and final manuscripts, ensuring consistency and adherence to the established standards of the ACL proceedings.

Key Specifications

The paper delineates numerous specifications for the authors to consider:

  1. Manuscript Format: Authors must use a two-column format on A4 paper, with precise specifications for margins, column width, and spacing between columns. This formatting is essential to maintain uniformity across all submissions.
  2. Submission Length and Structure:
    • Long papers can have up to 8 pages of content plus unlimited references, with an additional page provided in the final version to incorporate reviewers' comments.
    • Short papers are limited to 4 pages of content, also with unlimited pages for references. An additional page is allocated for the final versions.
  3. Anonymity: The reviewing process is double-blind, requiring authors to omit identifiable information in the initial submission, including author names and self-referential citations.
  4. Supplementary Material: Authors are encouraged to submit additional materials that aren't central to the paper but can aid in understanding or replicating the research. Such materials must comply with anonymity guidelines and must be self-contained.

Implications and Considerations

The standardization brought by these instructions is crucial for maintaining the integrity of the ACL conference proceedings. It ensures that every submission is evaluated based on its scientific merit, free of any formatting-related biases. Moreover, the guidelines foster transparency and replicability in computational linguistics research by encouraging the submission of supplementary materials, including code and data.

The paper highlights the importance of maintaining a strict adherence to the format to avoid rejection without review. This emphasizes the necessity for authors to meticulously follow the guidelines, underscoring the stringent standards ACL maintains for its conference publications.

Future Prospects

As research conferences evolve, the need for comprehensive and standardized submission guidelines remains vital. The instructions established in this paper could serve as a model for future conferences, reflecting a commitment to quality and accessibility. This foresight into structured submissions aids in the dissemination and utility of scholarly work across disciplines, reinforcing ACL's role as a leader in promoting rigorous scientific communication practices.

In the field of artificial intelligence, such clear directives assist authors in focusing on content quality and innovation. As AI research continues to advance, the protocols outlined here support the scalability and applicability of research findings, paving the way for ongoing contributions to the academic community.

In conclusion, the ACL 2020 instructions document is a pivotal resource for researchers aiming to contribute to the conference. Its detailed guidance ensures clarity, fairness, and consistency in the publication process, reflecting the conference’s high standards and dedication to advancing the field of computational linguistics.

Youtube Logo Streamline Icon: https://streamlinehq.com