Gmail Smart Compose

Chen, Mia Xu; Lee, Benjamin N.; Bansal, Gagan; Cao, Yuan; Zhang, Shuyuan; Lu, Justin; Tsay, Jackie; Wang, Yinan; Dai, Andrew M.; Chen, Zhifeng; Sohn, Timothy; Wu, Yonghui

doi:10.1145/3292500.3330723

Cited by 103 publications

(27 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Recently prediction service is applied on communication tools for composing email and chat messages to improve user writing productivity (Kannan et al, 2016;Deb et al, 2019;Chen et al, 2019;Microsoft Text Predictions, 2020).…”

Section: Related Workmentioning

confidence: 99%

“…In this work, we focus on the application of text prediction on production-level online communication tools, to help users compose emails (Chen et al, 2019;Microsoft Text Predictions, 2020), and in addition chat messages. In particular, we focus on examining useful contextual signals to give more accurate predicted text, using time, subject, and prior messages.…”

Section: Related Workmentioning

confidence: 99%

“…We combine these signals with the message body into a single "contextualized" string, using special tokens to separate signals, as shown in Figure 1a. This approach is inspired by (Chen et al, 2019), as they showed that concatenating contextual signals into a single input string gave a comparable result to more complex methods encoding these signals separately 1 . The remainder of this section explains details about each contextual signal we use.…”

Section: Contextualization Conceptsmentioning

confidence: 99%

“…The efficiency of these communications is enhanced by suggesting highly accurate text predictions with low latency. Text prediction services have been deployed across popular communication tools and platforms such as (Microsoft Text Predictions, 2020) or GMail Smart Compose (Chen et al, 2019).…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

When does text prediction benefit from additional context? An exploration of contextual signals for chat and email messages

Trajanovski¹,

Atalla²,

Kim³

et al. 2021

Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Langua

View full text Add to dashboard Cite

Email and chat communication tools are increasingly important for completing daily tasks. Accurate real-time phrase completion can save time and bolster productivity. Modern text prediction algorithms are based on large language models which typically rely on the prior words in a message to predict a completion. We examine how additional contextual signals (from previous messages, time, and subject) affect the performance of a commercial text prediction model. We compare contextual text prediction in chat and email messages from two of the largest commercial platforms Microsoft Teams and Outlook, finding that contextual signals contribute to performance differently between these scenarios. On emails, time context is most beneficial with small relative gains of 2% over baseline. Whereas, in chat scenarios, using a tailored set of previous messages as context yields relative improvements over the baseline between 9.3% and 18.6% across various critical serviceoriented text prediction metrics.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Contextualization Conceptsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

When does text prediction benefit from additional context? An exploration of contextual signals for chat and email messages

Trajanovski¹,

Atalla²,

Kim³

et al. 2021

Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Langua

View full text Add to dashboard Cite

show abstract

“…In SemEval 2014, van Gompel et al (2014) presented an L2 writing assistance task where systems find the proper translation of a word given a context in L2. Other writing assistance systems (not necessarily L2 learners) include assisting users with composing an email by auto-completion (Chen et al, 2019) and reply suggestion (Kannan et al, 2016).…”

Section: Related Workmentioning

confidence: 99%

TEASPN: Framework and Protocol for Integrated Writing Assistance Environments

Hagiwara¹,

Ito²,

Kuribayashi³

et al. 2019

Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conferen

View full text Add to dashboard Cite

Language technologies play a key role in assisting people with their writing. Although there has been steady progress in e.g., grammatical error correction (GEC), human writers are yet to benefit from this progress due to the high development cost of integrating with writing software. We propose TEASPN 1 , a protocol and an open-source framework for achieving integrated writing assistance environments. The protocol standardizes the way writing software communicates with servers that implement such technologies, allowing developers and researchers to integrate the latest developments in natural language processing (NLP) with low cost. As a result, users can enjoy the integrated experience in their favorite writing software. The results from experiments with human participants show that users use a wide range of technologies and rate their writing experience favorably, allowing them to write more fluent text.

show abstract

Evaluating Machine Unlearning: Applications, Approaches, and Accuracy

Ali,

Muhammad,

Adnan

et al. 2024

Engineering Reports

View full text Add to dashboard Cite

Machine learning (ML) enables computers to learn from experience by identifying patterns and trends. Despite ML's advancements in extracting valuable data, there are instances necessitating the removal or deletion of certain data, as ML models can inadvertently memorize training data. In many cases, ML models may memorize sensitive or personal data, raising concerns about data privacy and security. Machine unlearning (MU) techniques offer a solution to these concerns by selectively removing sensitive data from trained models without significantly compromising their performance. Similarly, we can analyze and evaluate whether MU can successfully achieve the “right to be forgotten.” In this paper, we investigate various MU approaches regarding their accuracy and potential applications. Experiments have shown that the data‐driven approach emerged as the most efficient method in terms of both time and accuracy, achieving a high level of precision with a minimal number of training epochs. When fine‐tuning, the full test error rises somewhat to 14.57% from the baseline model's 14.28%. One approach shows a high forget error of 99.90% with a full test error of 20.68%, while retraining yields a 100% forget error and a test error of 21.37%. While error‐minimizing noise preserves performance, the SCRUB technique results in a 21.08% test error and an 81.05% forget error, in contrast to the degradation brought on by error‐maximizing noise. On the other hand, the agnostic approach displayed sluggishness and generated less accurate results compared to the data‐driven approach. Furthermore, the choice of approach may depend on the unique requirements of the task and the available training resources.

show abstract

Gmail Smart Compose

Cited by 103 publications

References 36 publications

When does text prediction benefit from additional context? An exploration of contextual signals for chat and email messages

When does text prediction benefit from additional context? An exploration of contextual signals for chat and email messages

TEASPN: Framework and Protocol for Integrated Writing Assistance Environments

Evaluating Machine Unlearning: Applications, Approaches, and Accuracy

Contact Info

Product

Resources

About