Share This Author
Identifying content for planned events across social media sites
This paper focuses on the challenge of automatically identifying user-contributed content for events that are planned and, therefore, known in advance, across different social media sites, and develops query formulation strategies for retrieving content associated with an event on differentSocial media sites.
Automatic Detection of Incoherent Speech for Diagnosing Schizophrenia
A first benchmark comparison of previously proposed coherence models for detecting symptoms of schizophrenia are evaluated and evaluated on a new dataset of recorded interviews between subjects and clinicians and a novel computational model for reference incoherence based on ambiguous pronoun usage is proposed.
Pretraining with Contrastive Sentence Objectives Improves Discourse Performance of Language Models
- Dan Iter, Kelvin Guu, L. Lansing, Dan Jurafsky
- Computer ScienceAnnual Meeting of the Association for…
- 20 May 2020
Conpono, an inter-sentence objective for pretraining language models that models discourse coherence and the distance between sentences is proposed, and it is shown that Conpono yields gains of 2%-6% absolute even for tasks that do not explicitly evaluate discourse: textual entailment, common sense reasoning and reading comprehension.
From Laptop to Lambda: Outsourcing Everyday Jobs to Thousands of Transient Functional Containers
- Sadjad Fouladi, Francisco Romero, Keith Winstein
- Computer ScienceUSENIX Annual Technical Conference
- 10 July 2019
We present gg, a framework and a set of command-line tools that helps people execute everyday applications--e.g., software compilation, unit tests, video encoding, or object recognition--using…
Automatic Identification and Presentation of Twitter Content for Planned Events
A system for augmenting information about planned events with Twitter messages, using a set of automatic query building strategies, and two alternative interfaces to the system, namely, a browser plug-in and a customizable Web interface are presented.
Generating Adversarial Examples for Speech Recognition
This work shows successful results for two methods of generating adversarial examples where a high quality ASR system is fooled but the difference in the audio is imperceptible to the human ear.
A Thunk to Remember: make -j1000 (and other jobs) on functions-as-a-service infrastructure
- Sadjad Fouladi, Dan Iter, Shuvo Chatterjee, Christos Kozyrakis, M. Zaharia, Keith Winstein
- Computer Science
This work presents gg: a system for executing interdependent software workflows across thousands of short-lived “lambdas” that run in parallel on public cloud infrastructure with thousand-way parallelism, and finds that gg outperforms existing schemes for accelerating compilation.
Flipper: A Systematic Approach to Debugging Training Sets
The vision for Flipper is presented, a framework that presents users with high-level information about why their training set is inaccurate and informs their decisions as they improve their generative model manually, and potential tools within the Flipper framework are presented.
Socratic Learning: Augmenting Generative Models to Incorporate Latent Subsets in Training Data
Socratic learning is presented, a paradigm that uses feedback from a corresponding discriminative model to automatically identify subsets in the training data and augments the structure of the generative model accordingly, and shows that without any ground truth labels, the augmented generative models reduces error by up to 56.06% for a relation extraction task.
Generate rather than Retrieve: Large Language Models are Strong Context Generators
The proposed method is evaluated on three different knowledge-intensive tasks and its effectiveness on both zero-shot and supervised settings is demonstrated.