Machine learning (ML) models are only as good as the data they are trained on. But recent studies have found datasets widely used to train and evaluate ML models, e.g. ImageNet, to have pervasive labeling errors. Erroneous labels on the train set hurt ML models' ability to generalize, and they impact evaluation and model selection using the test set. Consequently, learning in the presence of labeling errors is an active area of research, yet this field lacks a comprehensive benchmark to evaluate these methods. Most of these methods are evaluated on a few computer vision datasets with significant variance in the experimental protocols. With such a large pool of methods and inconsistent evaluation, it is also unclear how ML practitioners can choose the right models to assess label quality in their data. To this end, we propose a benchmarking environment AQuA to rigorously evaluate methods that enable machine learning in the presence of label noise. We also introduce a design space to delineate concrete design choices of label error detection models. We hope that our proposed design space and benchmark enable practitioners to choose the right tools to improve their label quality and that our benchmark enables objective and rigorous evaluation of machine learning tools facing mislabeled data.

Computer network defence is a complicated task that has necessitated a high degree of human involvement. However, with recent advancements in machine learning, fully autonomous network defence is becoming increasingly plausible. This paper introduces an end-to-end methodology for studying attack strategies, designing defence agents and explaining their operation. First, using state diagrams, we visualise adversarial behaviour to gain insight about potential points of intervention and inform the design of our defensive models. We opt to use a set of deep reinforcement learning agents trained on different parts of the task and organised in a shallow hierarchy. Our evaluation shows that the resulting design achieves a substantial performance improvement compared to prior work. Finally, to better investigate the decision-making process of our agents, we complete our analysis with a feature ablation and importance study.

The wide applicability and adaptability of generative large language models (LLMs) has enabled their rapid adoption. While the pre-trained models can perform many tasks, such models are often fine-tuned to improve their performance on various downstream applications. However, this leads to issues over violation of model licenses, model theft, and copyright infringement. Moreover, recent advances show that generative technology is capable of producing harmful content which exacerbates the problems of accountability within model supply chains. Thus, we need a method to investigate how a model was trained or a piece of text was generated and what their pre-trained base model was. In this paper we take the first step to address this open problem by tracing back the origin of a given fine-tuned LLM to its corresponding pre-trained base model. We consider different knowledge levels and attribution strategies, and find that we can correctly trace back 8 out of the 10 fine tuned models with our best method.

As the capabilities of large language models (LLMs) continue to advance, evaluating their performance becomes increasingly crucial and challenging. This paper aims to bridge this gap by introducing CMMLU, a comprehensive Chinese benchmark that covers various subjects, including natural science, social sciences, engineering, and humanities. We conduct a thorough evaluation of 18 advanced multilingual- and Chinese-oriented LLMs, assessing their performance across different subjects and settings. The results reveal that most existing LLMs struggle to achieve an average accuracy of 50%, even when provided with in-context examples and chain-of-thought prompts, whereas the random baseline stands at 25%. This highlights significant room for improvement in LLMs. Additionally, we conduct extensive experiments to identify factors impacting the models' performance and propose directions for enhancing LLMs. CMMLU fills the gap in evaluating the knowledge and reasoning capabilities of large language models within the Chinese context.

In this study, we use recent stance detection methods to study the stance (for, against or neutral) of statements in official information booklets for voters. Our main goal is to answer the fundamental question: are topics to be voted on presented in a neutral way? To this end, we first train and compare several models for stance detection on a large dataset about Swiss politics. We find that fine-tuning an M-BERT model leads to the best accuracy. We then use our best model to analyze the stance of utterances extracted from the Swiss federal voting booklet concerning the Swiss popular votes of September 2022, which is the main goal of this project. We evaluated the models in both a multilingual as well as a monolingual context for German, French, and Italian. Our analysis shows that some issues are heavily favored while others are more balanced, and that the results are largely consistent across languages. Our findings have implications for the editorial process of future voting booklets and the design of better automated systems for analyzing political discourse.

The identification of compound-protein interactions (CPI) plays a critical role in drug screening, drug repurposing, and combination therapy studies. The effectiveness of CPI prediction relies heavily on the features extracted from both compounds and target proteins. While various prediction methods employ different feature combinations, both molecular-based and network-based models encounter the common obstacle of incomplete feature representations. Thus, a promising solution to this issue is to fully integrate all relevant CPI features. This study proposed a novel model named MCPI, which is designed to improve the prediction performance of CPI by integrating multiple sources of information, including the PPI network, CCI network, and structural features of CPI. The results of the study indicate that the MCPI model outperformed other existing methods for predicting CPI on public datasets. Furthermore, the study has practical implications for drug development, as the model was applied to search for potential inhibitors among FDA-approved drugs in response to the SARS-CoV-2 pandemic. The prediction results were then validated through the literature, suggesting that the MCPI model could be a useful tool for identifying potential drug candidates. Overall, this study has the potential to advance our understanding of CPI and guide drug development efforts.

Social media is awash with hateful content, much of which is often veiled with linguistic and topical diversity. The benchmark datasets used for hate speech detection do not account for such divagation as they are predominantly compiled using hate lexicons. However, capturing hate signals becomes challenging in neutrally-seeded malicious content. Thus, designing models and datasets that mimic the real-world variability of hate warrants further investigation. To this end, we present GOTHate, a large-scale code-mixed crowdsourced dataset of around 51k posts for hate speech detection from Twitter. GOTHate is neutrally seeded, encompassing different languages and topics. We conduct detailed comparisons of GOTHate with the existing hate speech datasets, highlighting its novelty. We benchmark it with 10 recent baselines. Our extensive empirical and benchmarking experiments suggest that GOTHate is hard to classify in a text-only setup. Thus, we investigate how adding endogenous signals enhances the hate speech detection task. We augment GOTHate with the user's timeline information and ego network, bringing the overall data source closer to the real-world setup for understanding hateful content. Our proposed solution HEN-mBERT is a modular, multilingual, mixture-of-experts model that enriches the linguistic subspace with latent endogenous signals from history, topology, and exemplars. HEN-mBERT transcends the best baseline by 2.5% and 5% in overall macro-F1 and hate class F1, respectively. Inspired by our experiments, in partnership with Wipro AI, we are developing a semi-automated pipeline to detect hateful content as a part of their mission to tackle online harm.

Extracting structured and grounded fact triples from raw text is a fundamental task in Information Extraction (IE). Existing IE datasets are typically collected from Wikipedia articles, using hyperlinks to link entities to the Wikidata knowledge base. However, models trained only on Wikipedia have limitations when applied to web domains, which often contain noisy text or text that does not have any factual information. We present WebIE, the first large-scale, entity-linked closed IE dataset consisting of 1.6M sentences automatically collected from the English Common Crawl corpus. WebIE also includes negative examples, i.e. sentences without fact triples, to better reflect the data on the web. We annotate ~21K triples from WebIE through crowdsourcing and introduce mWebIE, a translation of the annotated set in four other languages: French, Spanish, Portuguese, and Hindi. We evaluate the in-domain, out-of-domain, and zero-shot cross-lingual performance of generative IE models and find models trained on WebIE show better generalisability. We also propose three training strategies that use entity linking as an auxiliary task. Our experiments show that adding Entity-Linking objectives improves the faithfulness of our generative IE models.

Humans possess an extraordinary ability to create and utilize tools, allowing them to overcome physical limitations and explore new frontiers. With the advent of foundation models, AI systems have the potential to be equally adept in tool use as humans. This paradigm, i.e., tool learning with foundation models, combines the strengths of specialized tools and foundation models to achieve enhanced accuracy, efficiency, and automation in problem-solving. Despite its immense potential, there is still a lack of a comprehensive understanding of key challenges, opportunities, and future endeavors in this field. To this end, we present a systematic investigation of tool learning in this paper. We first introduce the background of tool learning, including its cognitive origins, the paradigm shift of foundation models, and the complementary roles of tools and models. Then we recapitulate existing tool learning research into tool-augmented and tool-oriented learning. We formulate a general tool learning framework: starting from understanding the user instruction, models should learn to decompose a complex task into several subtasks, dynamically adjust their plan through reasoning, and effectively conquer each sub-task by selecting appropriate tools. We also discuss how to train models for improved tool-use capabilities and facilitate the generalization in tool learning. Considering the lack of a systematic tool learning evaluation in prior works, we experiment with 18 representative tools and show the potential of current foundation models in skillfully utilizing tools. Finally, we discuss several open problems that require further investigation for tool learning. Overall, we hope this paper could inspire future research in integrating tools with foundation models.

Human behavior is conditioned by codes and norms that constrain action. Rules, ``manners,'' laws, and moral imperatives are examples of classes of constraints that govern human behavior. These systems of constraints are "messy:" individual constraints are often poorly defined, what constraints are relevant in a particular situation may be unknown or ambiguous, constraints interact and conflict with one another, and determining how to act within the bounds of the relevant constraints may be a significant challenge, especially when rapid decisions are needed. Despite such messiness, humans incorporate constraints in their decisions robustly and rapidly. General, artificially-intelligent agents must also be able to navigate the messiness of systems of real-world constraints in order to behave predictability and reliably. In this paper, we characterize sources of complexity in constraint processing for general agents and describe a computational-level analysis for such constraint compliance. We identify key algorithmic requirements based on the computational-level analysis and outline an initial, exploratory implementation of a general approach to constraint compliance.