Actionable Email Intent Modeling With Reparametrized RNNs
Authors: Chu-Cheng Lin, Dongyeop Kang, Michael Gamon, Patrick Pantel
AAAI 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | On a collection of datasets consisting of IRC, Reddit, and email, our reparametrized RNNs outperform common multitask/multidomain approaches on several speech act related tasks. We also experiment with a minimally supervised scenario of email recipient action classification, and find the reparametrized RNNs learn a useful representation. |
| Researcher Affiliation | Collaboration | Chu-Cheng Lin Johns Hopkins University Baltimore, MD clin103@jhu.edu, Dongyeop Kang Carnegie Mellon University Pittsburgh, PA dongyeok@cs.cmu.edu, Michael Gamon Microsoft Research Redmond, WA mgamon@microsoft.com, Patrick Pantel Microsoft Research Redmond, WA ppantel@microsoft.com. Madian Khabsa (madian@apple.com) and Ahmed Hassan Awadallah (hassanam@microsoft.com) are co-authors of this work. |
| Pseudocode | No | No pseudocode or algorithm blocks were found in the paper. Figure 1 shows computation graphs, not pseudocode. |
| Open Source Code | No | No explicit statement about releasing open-source code for the described methodology was found, nor a link to a code repository. The rebrand.ly link points to a corrected version of the paper, not code. |
| Open Datasets | Yes | In this study, all email messages we annotate and evaluate on are part of the Avocado dataset (Oard et al. 2015), which consists of emails and attachments taken from 279 accounts of a defunct information technology company referred to as Avocado .1 Email threads are reconstructed from the recipients mailboxes. We decided to use the Avocado dataset because it is the largest and newest one publicly available. The Ubuntu Dialog Corpus is a curated collection of chat logs from Ubuntu s Internet Relay Chat technical support channels (Lowe et al. 2015). |
| Dataset Splits | Yes | In all experiments in section 4, we use half of the dataset as training data, a quarter as the validation data and the remaining quarter as test data. |
| Hardware Specification | No | No specific hardware details (like GPU models, CPU types, or memory) used for running experiments were provided. |
| Software Dependencies | No | The paper mentions the use of the ADAM optimizer, but no specific software libraries or frameworks with version numbers are provided. |
| Experiment Setup | No | No specific hyperparameter values (e.g., learning rate, batch size, number of epochs, L2 regularization strength) or detailed system-level training settings were explicitly stated for the models. |