The rapid deployment of generative language models (LMs) has raised concerns about social biases affecting the well-being of diverse consumers. The extant literature on generative LMs has primarily examined bias via explicit identity prompting. However, prior research on bias in earlier language-based technology platforms, including search engines, has shown that discrimination can occur even when identity terms are not specified explicitly. Studies of bias in LM responses to open-ended prompts (where identity classifications are left unspecified) are lacking and have not yet been grounded in end-consumer harms. Here, we advance studies of generative LM bias by considering a broader set of natural use cases via open-ended prompting. In this "laissez-faire" setting, we find that synthetically generated texts from five of the most pervasive LMs (ChatGPT3.5, ChatGPT4, Claude2.0, Llama2, and PaLM2) perpetuate harms of omission, subordination, and stereotyping for minoritized individuals with intersectional race, gender, and/or sexual orientation identities (AI/AN, Asian, Black, Latine, MENA, NH/PI, Female, Non-binary, Queer). We find widespread evidence of bias to an extent that such individuals are hundreds to thousands of times more likely to encounter LM-generated outputs that portray their identities in a subordinated manner compared to representative or empowering portrayals. We also document a prevalence of stereotypes (e.g. perpetual foreigner) in LM-generated outputs that are known to trigger psychological harms that disproportionately affect minoritized individuals. These include stereotype threat, which leads to impaired cognitive performance and increased negative self-perception. Our findings highlight the urgent need to protect consumers from discriminatory harms caused by language models and invest in critical AI education programs tailored towards empowering diverse consumers.

We diagnose and briefly discuss the dearth of the author: a condition that arises when AI-based creativity support tools for writing allow users to produce large amounts of text without making a commensurate number of creative decisions, resulting in output that is sparse in expressive intent. We argue that the dearth of the author helps to explain a number of recurring difficulties and anxieties around AI-based writing support tools, but that it also suggests an ambitious new goal for AI-based CSTs.

Automatic prompt optimization is an important approach to improving the performance of large language models (LLMs). Recent research demonstrates the potential of using LLMs as prompt optimizers, which can generate improved task prompts via iterative refinement. In this paper, we propose a novel perspective to investigate the design of LLM-based prompt optimizers, by drawing an analogy with gradient-based model optimizers. To connect these two approaches, we identify two pivotal factors in model parameter learning: update direction and update method. Focused on the two aspects, we borrow the theoretical framework and learning methods from gradient-based optimization to design improved strategies for LLM-based prompt optimizers. By systematically analyzing a rich set of improvement strategies, we further develop a capable Gradient-inspired LLM-based Prompt Optimizer called GPO. At each step, it first retrieves relevant prompts from the optimization trajectory as the update direction. Then, it utilizes the generation-based refinement strategy to perform the update, while controlling the edit distance through a cosine-based decay strategy. Extensive experiments demonstrate the effectiveness and efficiency of GPO. In particular, GPO brings an additional improvement of up to 56.8% on Big-Bench Hard and 55.3% on MMLU compared to baseline methods.

Real-life graph data often expands continually, rendering the learning of graph neural networks (GNNs) on static graph data impractical. Graph continual learning (GCL) tackles this problem by continually adapting GNNs to the expanded graph of the current task while maintaining the performance over the graph of previous tasks. Memory replay-based methods, which aim to replay data of previous tasks when learning new tasks, have been explored as one principled approach to mitigate the forgetting of the knowledge learned from the previous tasks. In this paper we extend this methodology with a novel framework, called Debiased Lossless Memory replay (DeLoMe). Unlike existing methods that sample nodes/edges of previous graphs to construct the memory, DeLoMe learns small lossless synthetic node representations as the memory. The learned memory can not only preserve the graph data privacy but also capture the holistic graph information, for which the sampling-based methods are not viable. Further, prior methods suffer from bias toward the current task due to the data imbalance between the classes in the memory data and the current data. A debiased GCL loss function is devised in DeLoMe to effectively alleviate this bias. Extensive experiments on four graph datasets show the effectiveness of DeLoMe under both class- and task-incremental learning settings.

While human annotations play a crucial role in language technologies, annotator subjectivity has long been overlooked in data collection. Recent studies that have critically examined this issue are often situated in the Western context, and solely document differences across age, gender, or racial groups. As a result, NLP research on subjectivity have overlooked the fact that individuals within demographic groups may hold diverse values, which can influence their perceptions beyond their group norms. To effectively incorporate these considerations into NLP pipelines, we need datasets with extensive parallel annotations from various social and cultural groups. In this paper we introduce the \dataset dataset: a large-scale cross-cultural dataset of parallel annotations for offensive language in over 4.5K sentences annotated by a pool of over 4k annotators, balanced across gender and age, from across 21 countries, representing eight geo-cultural regions. The dataset contains annotators' moral values captured along six moral foundations: care, equality, proportionality, authority, loyalty, and purity. Our analyses reveal substantial regional variations in annotators' perceptions that are shaped by individual moral values, offering crucial insights for building pluralistic, culturally sensitive NLP models.

Task-oriented grasping (TOG), which refers to the problem of synthesizing grasps on an object that are configurationally compatible with the downstream manipulation task, is the first milestone towards tool manipulation. Analogous to the activation of two brain regions responsible for semantic and geometric reasoning during cognitive processes, modeling the complex relationship between objects, tasks, and grasps requires rich prior knowledge about objects and tasks. Existing methods typically limit the prior knowledge to a closed-set scope and cannot support the generalization to novel objects and tasks out of the training set. To address such a limitation, we propose FoundationGrasp, a foundation model-based TOG framework that leverages the open-ended knowledge from foundation models to learn generalizable TOG skills. Comprehensive experiments are conducted on the contributed Language and Vision Augmented TaskGrasp (LaViA-TaskGrasp) dataset, demonstrating the superiority of FoudationGrasp over existing methods when generalizing to novel object instances, object classes, and tasks out of the training set. Furthermore, the effectiveness of FoudationGrasp is validated in real-robot grasping and manipulation experiments on a 7 DoF robotic arm. Our code, data, appendix, and video are publicly available at https://sites.google.com/view/foundationgrasp.

A/B tests are often required to be conducted on subjects that might have social connections. For e.g., experiments on social media, or medical and social interventions to control the spread of an epidemic. In such settings, the SUTVA assumption for randomized-controlled trials is violated due to network interference, or spill-over effects, as treatments to group A can potentially also affect the control group B. When the underlying social network is known exactly, prior works have demonstrated how to conduct A/B tests adequately to estimate the global average treatment effect (GATE). However, in practice, it is often impossible to obtain knowledge about the exact underlying network. In this paper, we present UNITE: a novel estimator that relax this assumption and can identify GATE while only relying on knowledge of the superset of neighbors for any subject in the graph. Through theoretical analysis and extensive experiments, we show that the proposed approach performs better in comparison to standard estimators.

Dense passage retrieval (DPR) is the first step in the retrieval augmented generation (RAG) paradigm for improving the performance of large language models (LLM). DPR fine-tunes pre-trained networks to enhance the alignment of the embeddings between queries and relevant textual data. A deeper understanding of DPR fine-tuning will be required to fundamentally unlock the full potential of this approach. In this work, we explore DPR-trained models mechanistically by using a combination of probing, layer activation analysis, and model editing. Our experiments show that DPR training decentralizes how knowledge is stored in the network, creating multiple access pathways to the same information. We also uncover a limitation in this training style: the internal knowledge of the pre-trained model bounds what the retrieval model can retrieve. These findings suggest a few possible directions for dense retrieval: (1) expose the DPR training process to more knowledge so more can be decentralized, (2) inject facts as decentralized representations, (3) model and incorporate knowledge uncertainty in the retrieval process, and (4) directly map internal model knowledge to a knowledge base.

Specifically, we clarify the embedding structure of a DPP model in the exponential family of log-linear models (c.f., Agresti, 1990; Amari, 2001) in Theorem 1. Models embedded in exponential families are called curved exponential families. Information geometry (Amari, 1985) provides a measure, the e-embedding curvature tensor (Efron, 1975; Reeds, 1975; Amari, 1982; Sei, 2011), to quantify the extent to which a curved exponential family deviates from an exponential family. To check the e-embedding curvature as well as the Fisher information matrix, we apply the diagonal scaling (Marshall and Olkin, 1968), also known as the quality vs. diversity decomposition in the DPP literature (Kulesza and Taskar, 2012), to an L-ensemble kernel of a DPP model and then evaluate them, which clarifies that the subset of parameters related to the item-wise effects (quality terms) has zero e-embedding curvature (Corollary 1).

Relative Validity Indices (RVIs) such as the Silhouette Width Criterion, Calinski-Harabasz and Davie's Bouldin indices are the most popular tools for evaluating and optimising applications of clustering. Their ability to rank collections of candidate partitions has been used to guide the selection of the number of clusters, and to compare partitions from different clustering algorithms. Beyond these more conventional tasks, many examples can be found in the literature where RVIs have been used to compare and select other aspects of clustering approaches such as data normalisation procedures, data representation methods, and distance measures. The authors are not aware of any studies that have attempted to establish the suitability of RVIs for such comparisons. Moreover, given the impact of these aspects on pairwise similarities, it is not even immediately obvious how RVIs should be implemented when comparing these aspects. In this study, we conducted experiments with seven common RVIs on over 2.7 million clustering partitions for both synthetic and real-world datasets, encompassing feature-vector and time-series data. Our findings suggest that RVIs are not well-suited to these unconventional tasks, and that conclusions drawn from such applications may be misleading. It is recommended that normalisation procedures, representation methods, and distance measures instead be selected using external validation on high quality labelled datasets or carefully designed outcome-oriented objective criteria, both of which should be informed by relevant domain knowledge and clustering aims.