Shachi Paul

@ShachiPaul

NLU@Google, Past: Applied Scientist@Amazon, Student@LTICMU, Carnegie Mellon University. I work on Natural Language Understanding and Conversational Agents

California, USA
Vrijeme pridruživanja: travanj 2017.

Tweetovi

Blokirali ste korisnika/cu @ShachiPaul

Jeste li sigurni da želite vidjeti te tweetove? Time nećete deblokirati korisnika/cu @ShachiPaul

  1. proslijedio/la je Tweet
    19. pro 2019.

    The semantic parsing team in Google Assistant is looking for PhD interns in Mountain View. If you (or someone you know) want to work on projects involving spoken language understanding and few-shot learning please reach out to me. Kindly help retweet.

    Poništi
  2. proslijedio/la je Tweet
    25. stu 2019.

    We updated the MultiWOZ 2.1 release, now with user dialogue acts and slot descriptions included!

    Poništi
  3. proslijedio/la je Tweet
    28. lis 2019.

    The Schema-Guided Dialogue dataset, is the largest publicly available corpus of multi-domain, task-oriented dialogues for training the next generation of conversational digital assistants. Grab the dataset and learn about the schema-guided approach here ↓

    Poništi
  4. proslijedio/la je Tweet
    28. lis 2019.

    Super excited to be organizing the 2nd NLP For Conversational AI Workshop held at . If you like Seattle coffee ☕, Washington hikes ⛰️, and talking machines 🤖, we look forward to seeing you there!

    Prikaži ovu nit
    Poništi
  5. proslijedio/la je Tweet
    8. lis 2019.

    This is a nice diagram by Zhengyan Zhang and that shows how many recent pretrained language models are connected. The GitHub repo contains a full list of relevant papers:

    Poništi
  6. proslijedio/la je Tweet
    17. ruj 2019.

    We’ve released the Topical Chat Dataset, >230,000 turns of dialogue annotated with sources of factual claims. We believe it's the largest social-conversation and knowledge dataset available to the research community.

    Poništi
  7. proslijedio/la je Tweet
    16. ruj 2019.
    Poništi
  8. proslijedio/la je Tweet
    16. ruj 2019.

    If you are interested in task oriented dialogue systems check out the talk on hybrid dialogue state tracking (HyST) - Tuesday at 4:20 pm.

    Poništi
  9. proslijedio/la je Tweet
    16. ruj 2019.

    At ? Interested in building task-oriented dialogue systems? Check out the talk on work with and . Tuesday, 4 pm "Towards Universal Dialogue Act-Tagging for Task-Oriented Dialogues"

    Poništi
  10. 4. ruj 2019.

    Check out our work on dialogue act tagging and state tracking at Interspeech this year!

    Poništi
  11. proslijedio/la je Tweet
    4. ruj 2019.

    At , Alexa researchers report two new techniques for improving state tracking in task-oriented dialogue systems.

    Poništi
  12. proslijedio/la je Tweet
    19. srp 2019.

    Interesting interview with on this page outlining 's presence at in September. @AlexaTechWriter

    Poništi
  13. proslijedio/la je Tweet
    12. srp 2019.

    MultiWOZ 2.1 - we remove a lot of state annotation noise and rerun state of the art models for new baselines. Surprisingly the task becomes harder. Current SOTA (TRADE) -> 45.6 Joint goal accuracy

    Poništi
  14. proslijedio/la je Tweet
    9. srp 2019.

    This paper from scientists Shachi Paul published to today: Towards Universal Dialogue Act Tagging for Task-Oriented Dialogues

    Poništi
  15. proslijedio/la je Tweet
    2. srp 2019.

    HyST: A Hybrid Approach for Flexible and Accurate Dialogue State Tracking -- an Interspeech-approved paper from scientists @AlexaTechWriter

    Poništi
  16. proslijedio/la je Tweet
    5. lip 2019.

    We built the technology powering Alexa Conversations, which makes it possible for Alexa to plan a night out. Come talk to the Alexa AI team at re:MARS 2019. .

    Poništi
  17. proslijedio/la je Tweet
    5. lip 2019.

    At , Alexa VP and head scientist Rohit Prasad announced new AI technology that will enable Alexa to engage in multiturn dialogues that span skills, shifting the cognitive burden of complex task coordination from the customer to Alexa.

    Poništi
  18. proslijedio/la je Tweet
    5. lip 2019.

    Amazon announcing Conversations, a new deep learning based framework for modelling complex dialogue flows. Early skill builder reduced app complexity from 5k to 1.7k LoC.

    Prikaži ovu nit
    Poništi
  19. proslijedio/la je Tweet
    3. svi 2019.

    Interesting developments happened in 2018/2019 for natural language generation decoding algorithms: here's a thread with some papers & code So, the two most common decoders for language generation used to be greedy-decoding (GD) and beam-search (BS). [1/9]

    Prikaži ovu nit
    Poništi
  20. proslijedio/la je Tweet
    7. pro 2018.

    If you are interested in Dialog Systems, come to the workshop today. I will be presenting our poster on Fast and Flexible State Tracking.

    Poništi

Čini se da učitavanje traje već neko vrijeme.

Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.

    Možda bi vam se svidjelo i ovo:

    ·