In the evolving landscape of Environmental, Social, and Corporate Governance (ESG) impact assessment, the ML-ESG-2 shared task proposes identifying ESG impact types. To address this challenge, we present a comprehensive system leveraging ensemble learning techniques, capitalizing on early and late fusion approaches. Our approach employs four distinct models: mBERT, FlauBERT-base, ALBERT-base-v2, and a Multi-Layer Perceptron (MLP) incorporating Latent Semantic Analysis (LSA) and Term Frequency-Inverse Document Frequency (TF-IDF) features. Through extensive experimentation, we find that our early fusion ensemble approach, featuring the integration of LSA, TF-IDF, mBERT, FlauBERT-base, and ALBERT-base-v2, delivers the best performance. Our system offers a comprehensive ESG impact type identification solution, contributing to the responsible and sustainable decision-making processes vital in today's financial and corporate governance landscape.

Tool learning is widely acknowledged as a foundational approach or deploying large language models (LLMs) in real-world scenarios. While current research primarily emphasizes leveraging tools to augment LLMs, it frequently neglects emerging safety considerations tied to their application. To fill this gap, we present $ToolSword$, a comprehensive framework dedicated to meticulously investigating safety issues linked to LLMs in tool learning. Specifically, ToolSword delineates six safety scenarios for LLMs in tool learning, encompassing $malicious$ $queries$ and $jailbreak$ $attacks$ in the input stage, $noisy$ $misdirection$ and $risky$ $cues$ in the execution stage, and $harmful$ $feedback$ and $error$ $conflicts$ in the output stage. Experiments conducted on 11 open-source and closed-source LLMs reveal enduring safety challenges in tool learning, such as handling harmful queries, employing risky tools, and delivering detrimental feedback, which even GPT-4 is susceptible to. Moreover, we conduct further studies with the aim of fostering research on tool learning safety. The data is released in https://github.com/Junjie-Ye/ToolSword.

As machine learning (ML) becomes more prevalent in human-centric applications, there is a growing emphasis on algorithmic fairness and privacy protection. While previous research has explored these areas as separate objectives, there is a growing recognition of the complex relationship between privacy and fairness. However, previous works have primarily focused on examining the interplay between privacy and fairness through empirical investigations, with limited attention given to theoretical exploration. This study aims to bridge this gap by introducing a theoretical framework that enables a comprehensive examination of their interrelation. We shall develop and analyze an information bottleneck (IB) based information obfuscation method with local differential privacy (LDP) for fair representation learning. In contrast to many empirical studies on fairness in ML, we show that the incorporation of LDP randomizers during the encoding process can enhance the fairness of the learned representation. Our analysis will demonstrate that the disclosure of sensitive information is constrained by the privacy budget of the LDP randomizer, thereby enabling the optimization process within the IB framework to effectively suppress sensitive information while preserving the desired utility through obfuscation. Based on the proposed method, we further develop a variational representation encoding approach that simultaneously achieves fairness and LDP. Our variational encoding approach offers practical advantages. It is trained using a non-adversarial method and does not require the introduction of any variational prior. Extensive experiments will be presented to validate our theoretical results and demonstrate the ability of our proposed approach to achieve both LDP and fairness while preserving adequate utility.

Despite the success of reinforcement learning from human feedback (RLHF) in aligning language models with human values, reward hacking, also termed reward overoptimization, remains a critical challenge, which primarily stems from limitations in reward modeling, i.e., generalizability of the reward model and inconsistency in the preference dataset. In this work, we tackle this problem from an information theoretic-perspective, and propose a generalizable and robust framework for reward modeling, namely InfoRM, by introducing a variational information bottleneck objective to filter out irrelevant information and developing a mechanism for model complexity modulation. Notably, we further identify a correlation between overoptimization and outliers in the latent space, establishing InfoRM as a promising tool for detecting reward overoptimization. Inspired by this finding, we propose the Integrated Cluster Deviation Score (ICDS), which quantifies deviations in the latent space, as an indicator of reward overoptimization to facilitate the development of online mitigation strategies. Extensive experiments on a wide range of settings and model scales (70M, 440M, 1.4B, and 7B) support the effectiveness of InfoRM. Further analyses reveal that InfoRM's overoptimization detection mechanism is effective, potentially signifying a notable advancement in the field of RLHF. Code will be released upon acceptance.

Extensive efforts in automated approaches for content moderation have been focused on developing models to identify toxic, offensive, and hateful content with the aim of lightening the load for moderators. Yet, it remains uncertain whether improvements on those tasks have truly addressed moderators' needs in accomplishing their work. In this paper, we surface gaps between past research efforts that have aimed to provide automation for aspects of content moderation and the needs of volunteer content moderators, regarding identifying violations of various moderation rules. To do so, we conduct a model review on Hugging Face to reveal the availability of models to cover various moderation rules and guidelines from three exemplar forums. We further put state-of-the-art LLMs to the test, evaluating how well these models perform in flagging violations of platform rules from one particular forum. Finally, we conduct a user survey study with volunteer moderators to gain insight into their perspectives on useful moderation models. Overall, we observe a non-trivial gap, as missing developed models and LLMs exhibit moderate to low performance on a significant portion of the rules. Moderators' reports provide guides for future work on developing moderation assistant models.

The world is heading towards a dangerous place where selective government outrage and "a la carte" application of international law are becoming the norm. The result is already damning: a crisis of credibility and the erosion of trust in international institutions and governments, putting in peril the rules-based world order. As the heads of Amnesty International and Center for Civilians in Conflict, two of the world's most prominent organisations for human rights and protection of civilians, we have a simple demand for the world leaders who will be coming together on Friday for the 2024 Munich Security Conference: Protect international humanitarian and international human rights laws which are the best tools we have for protecting civilians in times of conflict, and stop creating exceptions that weaken rights protection and endanger global security and stability. Unfortunately, in 2023, world leaders responded unevenly to the countless violations of international humanitarian and human rights law we witnessed in various conflicts across the world. They expressed outrage at the crimes committed by some warring parties while offering diplomatic cover for others.

Although the charity would not name any specific sites where it has found the imagery, one popular AI art website has been accused of allowing users to publish sexual and graphic images of very young models. When the BBC approached Civit.ai about the issue in November, the firm said it takes potential CSAM on the site "very seriously" and asks the community to report images that users consider to "depict under-age characters/people in a mature or photorealistic context".

Machine learning models often excel on in-distribution (ID) data but struggle with unseen out-of-distribution (OOD) inputs. Most techniques for improving OOD robustness are not applicable to settings where the model is effectively a black box, such as when the weights are frozen, retraining is costly, or the model is leveraged via an API. Test-time augmentation (TTA) is a simple post-hoc technique for improving robustness that sidesteps black-box constraints by aggregating predictions across multiple augmentations of the test input. TTA has seen limited use in NLP due to the challenge of generating effective natural language augmentations. In this work, we propose LLM-TTA, which uses LLM-generated augmentations as TTA's augmentation function. LLM-TTA outperforms conventional augmentation functions across sentiment, toxicity, and news classification tasks for BERT and T5 models, with BERT's OOD robustness improving by an average of 4.30 percentage points without regressing average ID performance. We explore selectively augmenting inputs based on prediction entropy to reduce the rate of expensive LLM augmentations, allowing us to maintain performance gains while reducing the average number of generated augmentations by 57.76%. LLM-TTA is agnostic to the task model architecture, does not require OOD labels, and is effective across low and high-resource settings. We share our data, models, and code for reproducibility.

Selecting high-quality pre-training data is important for creating capable language models, but existing methods rely on simple heuristics. We introduce QuRating, a method for selecting pre-training data that captures the abstract qualities of texts which humans intuitively perceive. In this paper, we investigate four qualities - writing style, required expertise, facts & trivia, and educational value. We find that LLMs are able to discern these qualities and observe that they are better at making pairwise judgments of texts than at rating the quality of a text directly. We train a QuRater model to learn scalar ratings from pairwise judgments, and use it to annotate a 260B training corpus with quality ratings for each of the four criteria. In our experiments, we select 30B tokens according to the different quality ratings and train 1.3B-parameter language models on the selected data. We find that it is important to balance quality and diversity, as selecting only the highest-rated documents leads to poor results. When we sample using quality ratings as logits over documents, our models achieve lower perplexity and stronger in-context learning performance than baselines. Beyond data selection, we use the quality ratings to construct a training curriculum which improves performance without changing the training dataset. We extensively analyze the quality ratings and discuss their characteristics, biases, and wider implications.

Global ambient air pollution, a transboundary challenge, is typically addressed through interventions relying on data from spatially sparse and heterogeneously placed monitoring stations. These stations often encounter temporal data gaps due to issues such as power outages. In response, we have developed a scalable, data-driven, supervised machine learning framework. This model is designed to impute missing temporal and spatial measurements, thereby generating a comprehensive dataset for pollutants including NO$_2$, O$_3$, PM$_{10}$, PM$_{2.5}$, and SO$_2$. The dataset, with a fine granularity of 0.25$^{\circ}$ at hourly intervals and accompanied by prediction intervals for each estimate, caters to a wide range of stakeholders relying on outdoor air pollution data for downstream assessments. This enables more detailed studies. Additionally, the model's performance across various geographical locations is examined, providing insights and recommendations for strategic placement of future monitoring stations to further enhance the model's accuracy.