arXiv subjects

Tongshuang Wu

Publications and source records attributed to Tongshuang Wu.

At least 73 records · Page 4Linked to original sources

arXiv2023

Seeing Seeds Beyond Weeds: Green Teaming Generative AI for Beneficial Uses

Large generative AI models (GMs) like GPT and DALL-E are trained to generate content for general, wide-ranging purposes. GM content filters are generalized to filter out content which has a risk of harm in many cases, e.g., hate speech. However, prohibited content is not always harmful -- there are instances where generating prohibited content can be beneficial. So, when GMs filter out content, they preclude beneficial use cases along with harmful ones. Which use cases are precluded reflects the values embedded in GM content filtering. Recent work on red teaming proposes methods to bypass GM content filters to generate harmful content. We coin the term green teaming to describe methods of bypassing GM content filters to design for beneficial use cases. We showcase green teaming by: 1) Using ChatGPT as a virtual patient to simulate a person experiencing suicidal ideation, for suicide support training; 2) Using Codex to intentionally generate buggy solutions to train students on debugging; and 3) Examining an Instagram page using Midjourney to generate images of anti-LGBTQ+ politicians in drag. Finally, we discuss how our use cases demonstrate green teaming as both a practical design method and a mode of critique, which problematizes and subverts current understandings of harms and values in generative AI.

cs.HC↗

arXiv2023

BiasX: "Thinking Slow" in Toxic Content Moderation with Explanations of Implied Social Biases

Toxicity annotators and content moderators often default to mental shortcuts when making decisions. This can lead to subtle toxicity being missed, and seemingly toxic but harmless content being over-detected. We introduce BiasX, a framework that enhances content moderation setups with free-text explanations of statements' implied social biases, and explore its effectiveness through a large-scale crowdsourced user study. We show that indeed, participants substantially benefit from explanations for correctly identifying subtly (non-)toxic content. The quality of explanations is critical: imperfect machine-generated explanations (+2.4% on hard toxic examples) help less compared to expert-written human explanations (+7.2%). Our results showcase the promise of using free-text explanations to encourage more thoughtful toxicity moderation.

cs.CL↗

arXiv2023

Parachute: Evaluating Interactive Human-LM Co-writing Systems

A surge of advances in language models (LMs) has led to significant interest in using LMs to build co-writing systems, in which humans and LMs interactively contribute to a shared writing artifact. However, there is a lack of studies assessing co-writing systems in interactive settings. We propose a human-centered evaluation framework, Parachute, for interactive co-writing systems. Parachute showcases an integrative view of interaction evaluation, where each evaluation aspect consists of categorized practical metrics. Furthermore, we present Parachute with a use case to demonstrate how to evaluate and compare co-writing systems using Parachute.

cs.HC↗

arXiv2023

ScatterShot: Interactive In-context Example Curation for Text Transformation

The in-context learning capabilities of LLMs like GPT-3 allow annotators to customize an LLM to their specific tasks with a small number of examples. However, users tend to include only the most obvious patterns when crafting examples, resulting in underspecified in-context functions that fall short on unseen cases. Further, it is hard to know when "enough" examples have been included even for known patterns. In this work, we present ScatterShot, an interactive system for building high-quality demonstration sets for in-context learning. ScatterShot iteratively slices unlabeled data into task-specific patterns, samples informative inputs from underexplored or not-yet-saturated slices in an active learning manner, and helps users label more efficiently with the help of an LLM and the current example set. In simulation studies on two text perturbation scenarios, ScatterShot sampling improves the resulting few-shot functions by 4-5 percentage points over random sampling, with less variance as more examples are added. In a user study, ScatterShot greatly helps users in covering different patterns in the input space and labeling in-context examples more efficiently, resulting in better in-context learning and less user effort.

cs.HC↗

arXiv2023

Capabilities for Better ML Engineering

In spite of machine learning's rapid growth, its engineering support is scattered in many forms, and tends to favor certain engineering stages, stakeholders, and evaluation preferences. We envision a capability-based framework, which uses fine-grained specifications for ML model behaviors to unite existing efforts towards better ML engineering. We use concrete scenarios (model design, debugging, and maintenance) to articulate capabilities' broad applications across various different dimensions, and their impact on building safer, more generalizable and more trustworthy models that reflect human needs. Through preliminary experiments, we show capabilities' potential for reflecting model generalizability, which can provide guidance for ML engineering process. We discuss challenges and opportunities for capabilities' integration into ML engineering.

cs.AI↗

arXiv2022

Decisions that Explain Themselves: A User-Centric Deep Reinforcement Learning Explanation System

With deep reinforcement learning (RL) systems like autonomous driving being wildly deployed but remaining largely opaque, developers frequently use explainable RL (XRL) tools to better understand and work with deep RL agents. However, previous XRL works employ a techno-centric research approach, ignoring how RL developers perceive the generated explanations. Through a pilot study, we identify major goals for RL practitioners to use XRL methods and four pitfalls that widen the gap between existing XRL methods and these goals. The pitfalls include inaccessible reasoning processes, inconsistent or unintelligible explanations, and explanations that cannot be generalized. To fill the discovered gap, we propose a counterfactual-inference-based explanation method that discovers the details of the reasoning process of RL agents and generates natural language explanations. Surrounding this method, we build an interactive XRL system where users can actively explore explanations and influential information. In a user study with 14 participants, we validated that developers identified 20.9% more abnormal behaviors and limitations of RL agents with our system compared to the baseline method, and using our system helped end users improve their performance in actionability tests by 25.1% in an auto-driving task and by 16.9% in a StarCraft II micromanagement task.

cs.HC↗

arXiv2022

Towards Natural Language-Based Visualization Authoring

A key challenge to visualization authoring is the process of getting familiar with the complex user interfaces of authoring tools. Natural Language Interface (NLI) presents promising benefits due to its learnability and usability. However, supporting NLIs for authoring tools requires expertise in natural language processing, while existing NLIs are mostly designed for visual analytic workflow. In this paper, we propose an authoring-oriented NLI pipeline by introducing a structured representation of users' visualization editing intents, called editing actions, based on a formative study and an extensive survey on visualization construction tools. The editing actions are executable, and thus decouple natural language interpretation and visualization applications as an intermediate layer. We implement a deep learning-based NL interpreter to translate NL utterances into editing actions. The interpreter is reusable and extensible across authoring tools. The authoring tools only need to map the editing actions into tool-specific operations. To illustrate the usages of the NL interpreter, we implement an Excel chart editor and a proof-of-concept authoring tool, VisTalk. We conduct a user study with VisTalk to understand the usage patterns of NL-based authoring systems. Finally, we discuss observations on how users author charts with natural language, as well as implications for future research.

cs.HC↗

arXiv2022

NL-Augmenter: A Framework for Task-Sensitive Natural Language Augmentation

Data augmentation is an important component in the robustness evaluation of models in natural language processing (NLP) and in enhancing the diversity of the data they are trained on. In this paper, we present NL-Augmenter, a new participatory Python-based natural language augmentation framework which supports the creation of both transformations (modifications to the data) and filters (data splits according to specific features). We describe the framework and an initial set of 117 transformations and 23 filters for a variety of natural language tasks. We demonstrate the efficacy of NL-Augmenter by using several of its transformations to analyze the robustness of popular natural language models. The infrastructure, datacards and robustness analysis results are available publicly on the NL-Augmenter repository (https://github.com/GEM-benchmark/NL-Augmenter).

cs.CL↗

arXiv2022

It is AI's Turn to Ask Humans a Question: Question-Answer Pair Generation for Children's Story Books

Existing question answering (QA) techniques are created mainly to answer questions asked by humans. But in educational applications, teachers often need to decide what questions they should ask, in order to help students to improve their narrative understanding capabilities. We design an automated question-answer generation (QAG) system for this education scenario: given a story book at the kindergarten to eighth-grade level as input, our system can automatically generate QA pairs that are capable of testing a variety of dimensions of a student's comprehension skills. Our proposed QAG model architecture is demonstrated using a new expert-annotated FairytaleQA dataset, which has 278 child-friendly storybooks with 10,580 QA pairs. Automatic and human evaluations show that our model outperforms state-of-the-art QAG baseline systems. On top of our QAG system, we also start to build an interactive story-telling application for the future real-world deployment in this educational scenario.

cs.CL↗

arXiv2022

Fantastic Questions and Where to Find Them: FairytaleQA -- An Authentic Dataset for Narrative Comprehension

Question answering (QA) is a fundamental means to facilitate assessment and training of narrative comprehension skills for both machines and young children, yet there is scarcity of high-quality QA datasets carefully designed to serve this purpose. In particular, existing datasets rarely distinguish fine-grained reading skills, such as the understanding of varying narrative elements. Drawing on the reading education research, we introduce FairytaleQA, a dataset focusing on narrative comprehension of kindergarten to eighth-grade students. Generated by educational experts based on an evidence-based theoretical framework, FairytaleQA consists of 10,580 explicit and implicit questions derived from 278 children-friendly stories, covering seven types of narrative elements or relations. Our dataset is valuable in two folds: First, we ran existing QA models on our dataset and confirmed that this annotation helps assess models' fine-grained learning skills. Second, the dataset supports question generation (QG) task in the education domain. Through benchmarking with QG models, we show that the QG model trained on FairytaleQA is capable of asking high-quality and more diverse questions.

cs.CL↗

arXiv2022

Tailor: Generating and Perturbing Text with Semantic Controls

Controlled text perturbation is useful for evaluating and improving model generalizability. However, current techniques rely on training a model for every target perturbation, which is expensive and hard to generalize. We present Tailor, a semantically-controlled text generation system. Tailor builds on a pretrained seq2seq model and produces textual outputs conditioned on control codes derived from semantic representations. We craft a set of operations to modify the control codes, which in turn steer generation towards targeted attributes. These operations can be further composed into higher-level ones, allowing for flexible perturbation strategies. We demonstrate the effectiveness of these perturbations in multiple applications. First, we use Tailor to automatically create high-quality contrast sets for four distinct natural language processing (NLP) tasks. These contrast sets contain fewer spurious artifacts and are complementary to manually annotated ones in their lexical diversity. Second, we show that Tailor perturbations can improve model generalization through data augmentation. Perturbing just 2% of training data leads to a 5.8-point gain on an NLI challenge set measuring reliance on syntactic heuristics.

cs.CL↗

arXiv2022

AI Chains: Transparent and Controllable Human-AI Interaction by Chaining Large Language Model Prompts

Although large language models (LLMs) have demonstrated impressive potential on simple tasks, their breadth of scope, lack of transparency, and insufficient controllability can make them less effective when assisting humans on more complex tasks. In response, we introduce the concept of Chaining LLM steps together, where the output of one step becomes the input for the next, thus aggregating the gains per step. We first define a set of LLM primitive operations useful for Chain construction, then present an interactive system where users can modify these Chains, along with their intermediate results, in a modular way. In a 20-person user study, we found that Chaining not only improved the quality of task outcomes, but also significantly enhanced system transparency, controllability, and sense of collaboration. Additionally, we saw that users developed new ways of interacting with LLMs through Chains: they leveraged sub-tasks to calibrate model expectations, compared and contrasted alternative strategies by observing parallel downstream effects, and debugged unexpected model outputs by "unit-testing" sub-components of a Chain. In two case studies, we further explore how LLM Chains may be used in future applications

cs.HC↗

arXiv2022

Are Shortest Rationales the Best Explanations for Human Understanding?

Existing self-explaining models typically favor extracting the shortest possible rationales - snippets of an input text "responsible for" corresponding output - to explain the model prediction, with the assumption that shorter rationales are more intuitive to humans. However, this assumption has yet to be validated. Is the shortest rationale indeed the most human-understandable? To answer this question, we design a self-explaining model, LimitedInk, which allows users to extract rationales at any target length. Compared to existing baselines, LimitedInk achieves compatible end-task performance and human-annotated rationale agreement, making it a suitable representation of the recent class of self-explaining models. We use LimitedInk to conduct a user study on the impact of rationale length, where we ask human judges to predict the sentiment label of documents based only on LimitedInk-generated rationales with different lengths. We show rationales that are too short do not help humans predict labels better than randomly masked text, suggesting the need for more careful design of the best human rationales.

cs.CL↗

arXiv2022

StoryBuddy: A Human-AI Collaborative Chatbot for Parent-Child Interactive Storytelling with Flexible Parental Involvement

Despite its benefits for children's skill development and parent-child bonding, many parents do not often engage in interactive storytelling by having story-related dialogues with their child due to limited availability or challenges in coming up with appropriate questions. While recent advances made AI generation of questions from stories possible, the fully-automated approach excludes parent involvement, disregards educational goals, and underoptimizes for child engagement. Informed by need-finding interviews and participatory design (PD) results, we developed StoryBuddy, an AI-enabled system for parents to create interactive storytelling experiences. StoryBuddy's design highlighted the need for accommodating dynamic user needs between the desire for parent involvement and parent-child bonding and the goal of minimizing parent intervention when busy. The PD revealed varied assessment and educational goals of parents, which StoryBuddy addressed by supporting configuring question types and tracking child progress. A user study validated StoryBuddy's usability and suggested design insights for future parent-AI collaboration systems.

cs.HC↗

arXiv2022

PromptChainer: Chaining Large Language Model Prompts through Visual Programming

While LLMs can effectively help prototype single ML functionalities, many real-world applications involve complex tasks that cannot be easily handled via a single run of an LLM. Recent work has found that chaining multiple LLM runs together (with the output of one step being the input to the next) can help users accomplish these more complex tasks, and in a way that is perceived to be more transparent and controllable. However, it remains unknown what users need when authoring their own LLM chains -- a key step for lowering the barriers for non-AI-experts to prototype AI-infused applications. In this work, we explore the LLM chain authoring process. We conclude from pilot studies find that chaining requires careful scaffolding for transforming intermediate node outputs, as well as debugging the chain at multiple granularities; to help with these needs, we designed PromptChainer, an interactive interface for visually programming chains. Through case studies with four people, we show that PromptChainer supports building prototypes for a range of applications, and conclude with open questions on scaling chains to complex tasks, and supporting low-fi chain prototyping.

cs.HC↗

arXiv2021

Pretty Princess vs. Successful Leader: Gender Roles in Greeting Card Messages

People write personalized greeting cards on various occasions. While prior work has studied gender roles in greeting card messages, systematic analysis at scale and tools for raising the awareness of gender stereotyping remain under-investigated. To this end, we collect a large greeting card message corpus covering three different occasions (birthday, Valentine's Day and wedding) from three sources (exemplars from greeting message websites, real-life greetings from social media and language model generated ones). We uncover a wide range of gender stereotypes in this corpus via topic modeling, odds ratio and Word Embedding Association Test (WEAT). We further conduct a survey to understand people's perception of gender roles in messages from this corpus and if gender stereotyping is a concern. The results show that people want to be aware of gender roles in the messages, but remain unconcerned unless the perceived gender roles conflict with the recipient's true personality. In response, we developed GreetA, an interactive visualization and writing assistant tool to visualize fine-grained topics in greeting card messages drafted by the users and the associated gender perception scores, but without suggesting text changes as an intervention.

cs.HC↗

arXiv2021

DeHumor: Visual Analytics for Decomposing Humor

Despite being a critical communication skill, grasping humor is challenging -- a successful use of humor requires a mixture of both engaging content build-up and an appropriate vocal delivery (e.g., pause). Prior studies on computational humor emphasize the textual and audio features immediately next to the punchline, yet overlooking longer-term context setup. Moreover, the theories are usually too abstract for understanding each concrete humor snippet. To fill in the gap, we develop DeHumor, a visual analytical system for analyzing humorous behaviors in public speaking. To intuitively reveal the building blocks of each concrete example, DeHumor decomposes each humorous video into multimodal features and provides inline annotations of them on the video script. In particular, to better capture the build-ups, we introduce content repetition as a complement to features introduced in theories of computational humor and visualize them in a context linking graph. To help users locate the punchlines that have the desired features to learn, we summarize the content (with keywords) and humor feature statistics on an augmented time matrix. With case studies on stand-up comedy shows and TED talks, we show that DeHumor is able to highlight various building blocks of humor examples. In addition, expert interviews with communication coaches and humor researchers demonstrate the effectiveness of DeHumor for multimodal humor analysis of speech content and vocal delivery.

cs.CL↗

arXiv2021

Polyjuice: Generating Counterfactuals for Explaining, Evaluating, and Improving Models

While counterfactual examples are useful for analysis and training of NLP models, current generation methods either rely on manual labor to create very few counterfactuals, or only instantiate limited types of perturbations such as paraphrases or word substitutions. We present Polyjuice, a general-purpose counterfactual generator that allows for control over perturbation types and locations, trained by finetuning GPT-2 on multiple datasets of paired sentences. We show that Polyjuice produces diverse sets of realistic counterfactuals, which in turn are useful in various distinct applications: improving training and evaluation on three different tasks (with around 70% less annotation effort than manual generation), augmenting state-of-the-art explanation techniques, and supporting systematic counterfactual error analysis by revealing behaviors easily missed by human experts.

cs.CL↗