LLMs-as-a-judge is a recently popularized method which replaces human judgements in task evaluation with automatic evaluation using LLMs. Due to widespread use of RLHF (Reinforcement Learning from Human Feedback), state-of-the-art LLMs like GPT4 and Llama3 are expected to have strong alignment with human preferences when prompted for a quality judgement, such as the coherence of a text. While this seems beneficial, it is not clear whether the assessments by an LLM-as-a-judge constitute only an evaluation based on the instructions in the prompts, or reflect its preference for high-quality data similar to its fine-tune data. To investigate how much influence prompting the LLMs-as-a-judge has on the alignment of AI judgements to human judgements, we analyze prompts with increasing levels of instructions about the target quality of an evaluation, for several LLMs-as-a-judge. Further, we compare to a prompt-free method using model perplexity as a quality measure instead. We aggregate a taxonomy of quality criteria commonly used across state-of-the-art evaluations with LLMs and provide this as a rigorous benchmark of models as judges. Overall, we show that the LLMs-as-a-judge benefit only little from highly detailed instructions in prompts and that perplexity can sometimes align better with human judgements than prompting, especially on textual quality.

String data is common in real-world datasets: 67.6% of values in a sample of 1.8 million real Excel spreadsheets from the web were represented as text. Automatically cleaning such string data can have a significant impact on users. Previous approaches are limited to error detection, require that the user provides annotations, examples, or constraints to fix the errors, and focus independently on syntactic errors or semantic errors in strings, but ignore that strings often contain both syntactic and semantic substrings. We introduce DataVinci, a fully unsupervised string data error detection and repair system. DataVinci learns regular-expression-based patterns that cover a majority of values in a column and reports values that do not satisfy such majority patterns as data errors. DataVinci can automatically derive edits to the data error based on the majority patterns and using row tuples associated with majority values as examples. To handle strings with both syntactic and semantic substrings, DataVinci uses an LLM to abstract (and re-concretize) portions of strings that are semantic. Because not all data columns can result in majority patterns, when available, DataVinci can leverage execution information from an existing data program (which uses the target data as input) to identify and correct data repairs that would not otherwise be identified. DataVinci outperforms eleven baseline systems on both data error detection and repair as demonstrated on four existing and new benchmarks.

LLM-powered tools like ChatGPT Data Analysis, have the potential to help users tackle the challenging task of data analysis programming, which requires expertise in data processing, programming, and statistics. However, our formative study (n=15) uncovered serious challenges in verifying AI-generated results and steering the AI (i.e., guiding the AI system to produce the desired output). We developed two contrasting approaches to address these challenges. The first (Stepwise) decomposes the problem into step-by-step subgoals with pairs of editable assumptions and code until task completion, while the second (Phasewise) decomposes the entire problem into three editable, logical phases: structured input/output assumptions, execution plan, and code. A controlled, within-subjects experiment (n=18) compared these systems against a conversational baseline. Users reported significantly greater control with the Stepwise and Phasewise systems, and found intervention, correction, and verification easier, compared to the baseline. The results suggest design guidelines and trade-offs for AI-assisted data analysis tools.

Code-generating large language models map natural language to code. However, only a small portion of the infinite space of naturalistic utterances is effective at guiding code generation. For non-expert end-user programmers, learning this is the challenge of abstraction matching. We examine this challenge in the specific context of data analysis in spreadsheets, in a system that maps the user’s natural language query to Python code using the Codex generator, executes the code, and shows the result. We propose grounded abstraction matching, which bridges the abstraction gap by translating the code back into a systematic and predictable naturalistic utterance. In a between-subjects, think-aloud study (n=24), we compare grounded abstraction matching to an ungrounded alternative based on previously established query framing principles. We find that the grounded approach improves end-users’ understanding of the scope and capabilities of the code-generating model, and the kind of language needed to use it effectively.

Imagine a developer who can only change their last line of code-how often would they have to start writing a function from scratch before it is correct?Auto-regressive models for code generation from natural language have a similar limitation: they do not easily allow reconsidering earlier tokens generated.We introduce CODEFUSION, a pre-trained diffusion code generation model that addresses this limitation by iteratively denoising a complete program conditioned on the encoded natural language.We evaluate CODEFUSION on the task of natural language to code generation for Bash, Python, and Microsoft Excel conditional formatting (CF) rules.Experiments show that CODEFU-SION (75M parameters) performs on par with state-of-the-art auto-regressive systems (350M-175B parameters) in top-1 accuracy and outperforms them in top-3 and top-5 accuracy, due to its better balance in diversity versus quality.

Code-generating large language models (LLMs) are transforming programming. Their capability to generate multi-step solutions provides even non-programmers a mechanism to harness the power of coding. Non-programmers often use spreadsheets to manage tabular data, as they offer an intuitive understanding of data manipulation and formula out-comes. Considering that LLMs can generate complex, potentially incorrect code, our focus is on enabling user trust in the accuracy of LLM-generated code. We present ColDeco, the first end-user inspection tool for comprehending code produced by LLMs for tabular data tasks. ColDeco integrates two new features for inspection with a grid-based interface. First, users can decompose a generated solution into intermediate helper columns to understand how the problem is solved step by step. Second, users can interact with a filtered table of summary rows, which highlight interesting cases in the program. We evaluate our tool using a within-subjects user study (n=24) where participants are asked to verify the correctness of programs generated by an LLM. We found that while all features are independently useful, participants preferred them in combination. Users especially noted the usefulness of helper columns, but wanted more transparency in how summary rows are generated to assist with understanding and trusting them. Users also highlighted the application of ColDeco in collaborative settings for explaining and understanding existing formulas.

Formatting is an important property in tables for visualization, presentation, and analysis. Spreadsheet software allows users to automatically format their tables by writing data-dependent conditional formatting (CF) rules. Writing such rules is often challenging for users as it requires understanding and implementing the underlying logic. We present FormaT5, a transformer-based model that can generate a CF rule given the target table and a natural language description of the desired formatting logic. We find that user descriptions for these tasks are often under-specified or ambiguous, making it harder for code generation systems to accurately learn the desired rule in a single step. To tackle this problem of under-specification and minimise argument errors, FormaT5 learns to predict placeholders though an abstention objective. These placeholders can then be filled by a second model or, when examples of rows that should be formatted are available, by a programming-by-example system. To evaluate FormaT5 on diverse and real scenarios, we create an extensive benchmark of 1053 CF tasks, containing real-world descriptions collected from four different sources. We release our benchmarks to encourage research in this area. Abstention and filling allow FormaT5 to outperform 8 different neural approaches on our benchmarks, both with and without examples. Our results illustrate the value of building domain-specific learning systems.

Spreadsheets are widely used for table manipulation and presentation. Stylistic formatting of these tables is an important property for presentation and analysis. As a result, popular spreadsheet software, such as Excel, supports automatically formatting tables based on rules. Unfortunately, writing such formatting rules can be challenging for users as it requires knowledge of the underlying rule language and data logic. We present Cornet, a system that tackles the novel problem of automatically learning such formatting rules from user-provided formatted cells. Cornet takes inspiration from advances in inductive programming and combines symbolic rule enumeration with a neural ranker to learn conditional formatting rules. To motivate and evaluate our approach, we extracted tables with over 450K unique formatting rules from a corpus of over 1.8M real worksheets. Since we are the first to introduce the task of automatically learning conditional formatting rules, we compare Cornet to a wide range of symbolic and neural baselines adapted from related domains. Our results show that Cornet accurately learns rules across varying setups. Additionally, we show that in some cases Cornet can find rules that are shorter than those written by users and can also discover rules in spreadsheets that users have manually formatted. Furthermore, we present two case studies investigating the generality of our approach by extending Cornet to related data tasks (e.g., filtering) and generalizing to conditional formatting over multiple columns.

Data management and analysis tasks are often carried out using spreadsheet software. A popular feature in most spreadsheet platforms is the ability to define data-dependent formatting rules. These rules can express actions such as "color red all entries in a column that are negative" or "bold all rows not containing error or failure". Unfortunately, users who want to exercise this functionality need to manually write these conditional formatting (CF) rules. We introduce Cornet, a system that automatically learns such conditional formatting rules from user examples. Cornet takes inspiration from inductive program synthesis and combines symbolic rule enumeration, based on semi-supervised clustering and iterative decision tree learning, with a neural ranker to produce accurate conditional formatting rules. In this demonstration, we show Cornet in action as a simple add-in to Microsoft's Excel. After the user provides one or two formatted cells as examples, Cornet generates formatting rule suggestions for the user to apply to the spreadsheet.

The following topics are dealt with: computer science education; programming; software tools; computer aided instruction; software engineering; interactive systems; learning (artificial intelligence); data analysis; text analysis; groupware.