Charlie Welch

Philipps University of Marburg | PUM

Large pre-trained neural language models have supported the effectiveness of many NLP tasks, yet are still prone to generating toxic language hindering the safety of their use. Using empathetic data, we improve over recent work on controllable text generation that aims to reduce the toxicity of generated text. We find we are able to dramatically re...

We introduce the problem of proficiency modeling: Given a user's posts on a social media platform, the task is to identify the subset of posts or topics for which the user has some level of proficiency. This enables the filtering and ranking of social media posts on a given topic as per user proficiency. Unlike experts on a given topic, proficient...

A growing number of people engage in online health forums, making it important to understand the quality of the advice they receive. In this paper, we explore the role of expertise in responses provided to help-seeking posts regarding mental health. We study the differences between (1) interactions with peers; and (2) interactions with self-identif...

This thesis focuses on personalization, word representations, and longitudinal dialog. We first look at users expressions of individual preferences. In this targeted sentiment task, we find that we can improve entity extraction and sentiment classification using domain lexicons and linear term weighting. This task is important to personalization an...

In this paper, we introduce personalized word embeddings, and examine their value for language modeling. We compare the performance of our proposed prediction model when using personalized versus generic word representations, and study how these representations can be leveraged for improved performance. We provide insight into what types of words c...

Word embeddings are usually derived from corpora containing text from many individuals, thus leading to general purpose representations rather than individually personalized representations. While personalized embeddings can be useful to improve language model performance and other language processing tasks, they can only be computed for people wit...

Many NLP applications, such as biomedical data and technical support, have 10-100 million tokens of in-domain data and limited computational resources for learning from it. How should we train a language model in this scenario? Most language modeling research considers either a small dataset with a closed vocabulary (like the standard 1 million tok...

The ongoing COVID-19 pandemic has raised concerns for many regarding personal and public health implications, financial security and economic stability. Alongside many other unprecedented challenges, there are increasing concerns over social isolation and mental health. We introduce \textit{Expressive Interviewing}--an interview-style conversationa...

We explore the use of longitudinal dialog data for two dialog prediction tasks: next message prediction and response time prediction. We show that a neural model using personal data that leverages a combination of message content, style matching, time features, and speaker attributes leads to the best results for both tasks, with error rate reducti...

We examine a large dialog corpus obtained from the conversation history of a single individual with 104 conversation partners. The corpus consists of half a million instant messages, across several messaging platforms. We focus our analyses on seven speaker attributes, each of which partitions the set of speakers, namely: gender; relative age; fami...

Understanding current world events in real-time involves sifting through news articles, tweets, photos, and videos from many different perspectives. The goal of the DARPA-funded AIDA project is to automate much of this process, building a knowledge base that can be queried to strategically generate hypotheses about different aspects of an event. We...