ABSTRACT As governments race to implement new climate adaptation policies that prepare for more frequent flooding, they must seek policies that are effective for all communities and uphold climate justice. This requires evaluating policies not only on their overall effectiveness but also on whether their benefits are felt across all communities. We illustrate the importance of considering such disparities for flood adaptation using the FEMA National Flood Insurance Program Community Rating System and its dataset of 2.5 million flood insurance claims. We use CausalFlow, a causal inference method based on deep generative models, to estimate the treatment effect of flood adaptation interventions based on a community's income, diversity, population, flood risk, educational attainment, and precipitation. We find that the program saves communities $5,000-15,000 per household. However, these savings are not evenly spread across communities. For example, for low-income communities savings sharply decline as flood-risk increases in contrast to their high-income counterparts with all else equal. Even among low-income communities, there is a gap in savings between predominantly white and non-white communities: savings of predominantly white communities can be higher by more than $6000 per household. As communities worldwide ramp up efforts to reduce losses inflicted by floods, simply prescribing a series flood adaptation measures is not enough. Programs must provide communities with the necessary technical and economic support to compensate for historical patterns of disenfranchisement, racism, and inequality. Future flood adaptation efforts should go beyond reducing losses overall and aim to close existing gaps to equitably support communities in the race for climate adaptation. INTRODUCTION Flooding constitutes nearly a third of all losses from natural disasters worldwide (Reuters 2022). By the end of the century, rising sea levels and coastal flooding are estimated to cost the global economy $14.2 trillion (a fifth of the global GDP) in damaged assets (Kirezci et al. 2020).

As language models (LMs) advance, interest is growing in applying them to high-stakes societal decisions, such as determining financing or housing eligibility. However, their potential for discrimination in such contexts raises ethical concerns, motivating the need for better methods to evaluate these risks. We present a method for proactively evaluating the potential discriminatory impact of LMs in a wide range of use cases, including hypothetical use cases where they have not yet been deployed. Specifically, we use an LM to generate a wide array of potential prompts that decision-makers may input into an LM, spanning 70 diverse decision scenarios across society, and systematically vary the demographic information in each prompt. Applying this methodology reveals patterns of both positive and negative discrimination in the Claude 2.0 model in select settings when no interventions are applied. While we do not endorse or permit the use of language models to make automated decisions for the high-risk use cases we study, we demonstrate techniques to significantly decrease both positive and negative discrimination through careful prompt engineering, providing pathways toward safer deployment in use cases where they may be appropriate. Our work enables developers and policymakers to anticipate, measure, and address discrimination as language model capabilities and applications continue to expand. We release our dataset and prompts at https://huggingface.co/datasets/Anthropic/discrim-eval

The "Reversal Curse" refers to the scenario where auto-regressive decoder large language models (LLMs), such as ChatGPT, trained on "A is B" fail to learn "B is A", demonstrating a basic failure of logical deduction. This raises a red flag in the use of GPT models for certain general tasks such as constructing knowledge graphs, considering their adherence to this symmetric principle. In our study, we examined a bidirectional LLM, BERT, and found that it is immune to the reversal curse. Driven by ongoing efforts to construct biomedical knowledge graphs with LLMs, we also embarked on evaluating more complex but essential deductive reasoning capabilities. This process included first training encoder and decoder language models to master the intersection ($\cap$) and union ($\cup$) operations on two sets and then moving on to assess their capability to infer different combinations of union ($\cup$) and intersection ($\cap$) operations on three newly created sets. The findings showed that while both encoder and decoder language models, trained for tasks involving two sets (union/intersection), were proficient in such scenarios, they encountered difficulties when dealing with operations that included three sets (various combinations of union and intersection). Our research highlights the distinct characteristics of encoder and decoder models in simple and complex logical reasoning. In practice, the choice between BERT and GPT should be guided by the specific requirements and nature of the task at hand, leveraging their respective strengths in bidirectional context comprehension and sequence prediction.

Deep learning has made notable progress across a diverse range of machine learning domains, including image classification, object detection and speech recognition. The influence of deep learning extends to an increasing number of real-world applications and systems. However, recent research highlights the susceptibility of deep learning models to well-designed input samples, called adversarial examples. Adversarial examples are imperceptible to humans but can easily fool deep neural networks [1]. Szedegy et al. [2] manipulated state-of-the-art deep neural networks to misclassify images by applying subtle and barely perceptible perturbations, as seen in Figure 1, showcasing the vulnerability of these models to targeted errors. Researchers discovered that a single perturbation could lead a neural network to misclassify multiple classes, revealing inherent weaknesses in the training algorithms. Despite neural networks being nonlinear functions [3], it was surprising to find that linear approximations are effective in creating adversarial perturbations. This unexpected result is attributed to the design of easily trainable models, which empirically exhibit an accurate local linear approximation for their loss function [4], allowing adversaries to make subtle changes in the pixel values that lead to misclassification. Figure 1: Szegedy et al. [2] were able to fool AlexNet [5] by classifying a perturbed image of a dog into an ostrich There has been a good amount of progress in research pertaining to defending against these adversarial attacks. The defenses aim to either detect whether an input is adversarial, or aim to modify the input so that it is no longer adversarial in nature.

We present the Granular AMR Parsing Evaluation Suite (GrAPES), a challenge set for Abstract Meaning Representation (AMR) parsing with accompanying evaluation metrics. AMR parsers now obtain high scores on the standard AMR evaluation metric Smatch, close to or even above reported inter-annotator agreement. But that does not mean that AMR parsing is solved; in fact, human evaluation in previous work indicates that current parsers still quite frequently make errors on node labels or graph structure that substantially distort sentence meaning. Here, we provide an evaluation suite that tests AMR parsers on a range of phenomena of practical, technical, and linguistic interest. Our 36 categories range from seen and unseen labels, to structural generalization, to coreference. GrAPES reveals in depth the abilities and shortcomings of current AMR parsers.

The challenge of teaching robots to perform dexterous manipulation, dynamic locomotion, or whole--body manipulation from a small number of demonstrations is an important research field that has attracted interest from across the robotics community. In this work, we propose a novel approach by joining the theories of Koopman Operators and Dynamic Movement Primitives to Learning from Demonstration. Our approach, named \gls{admd}, projects nonlinear dynamical systems into linear latent spaces such that a solution reproduces the desired complex motion. Use of an autoencoder in our approach enables generalizability and scalability, while the constraint to a linear system attains interpretability. Our results are comparable to the Extended Dynamic Mode Decomposition on the LASA Handwriting dataset but with training on only a small fractions of the letters.

Previous studies regarding the perception of emotions for embodied virtual agents have shown the effectiveness of using virtual characters in conveying emotions through interactions with humans. However, creating an autonomous embodied conversational agent with expressive behaviors presents two major challenges. The first challenge is the difficulty of synthesizing the conversational behaviors for each modality that are as expressive as real human behaviors. The second challenge is that the affects are modeled independently, which makes it difficult to generate multimodal responses with consistent emotions across all modalities. In this work, we propose a conceptual framework, ACTOR (Affect-Consistent mulTimodal behaviOR generation), that aims to increase the perception of affects by generating multimodal behaviors conditioned on a consistent driving affect. We have conducted a user study with 199 participants to assess how the average person judges the affects perceived from multimodal behaviors that are consistent and inconsistent with respect to a driving affect. The result shows that among all model conditions, our affect-consistent framework receives the highest Likert scores for the perception of driving affects. Our statistical analysis suggests that making a modality affect-inconsistent significantly decreases the perception of driving affects. We also observe that multimodal behaviors conditioned on consistent affects are more expressive compared to behaviors with inconsistent affects. Therefore, we conclude that multimodal emotion conditioning and affect consistency are vital to enhancing the perception of affects for embodied conversational agents.

We introduce SwiftSage, a novel agent framework inspired by the dual-process theory of human cognition, designed to excel in action planning for complex interactive reasoning tasks. SwiftSage integrates the strengths of behavior cloning and prompting large language models (LLMs) to enhance task completion performance. The framework comprises two primary modules: the Swift module, representing fast and intuitive thinking, and the Sage module, emulating deliberate thought processes. The Swift module is a small encoder-decoder LM fine-tuned on the oracle agent's action trajectories, while the Sage module employs LLMs such as GPT-4 for subgoal planning and grounding. We develop a heuristic method to harmoniously integrate the two modules, resulting in a more efficient and robust problem-solving process. In 30 tasks from the ScienceWorld benchmark, SwiftSage significantly outperforms other methods such as SayCan, ReAct, and Reflexion, demonstrating its effectiveness in solving complex interactive tasks.

General-purpose knowledge bases (KBs) are a cornerstone of knowledge-centric AI. Many of them are constructed pragmatically from Web sources, and are thus far from complete. This poses challenges for the consumption as well as the curation of their content. While several surveys target the problem of completing incomplete KBs, the first problem is arguably to know whether and where the KB is incomplete in the first place, and to which degree. In this survey we discuss how knowledge about completeness, recall, and negation in KBs can be expressed, extracted, and inferred. We cover (i) the logical foundations of knowledge representation and querying under partial closed-world semantics; (ii) the estimation of this information via statistical patterns; (iii) the extraction of information about recall from KBs and text; (iv) the identification of interesting negative statements; and (v) relaxed notions of relative recall. This survey is targeted at two types of audiences: (1) practitioners who are interested in tracking KB quality, focusing extraction efforts, and building quality-aware downstream applications; and (2) data management, knowledge base and semantic web researchers who wish to understand the state of the art of knowledge bases beyond the open-world assumption. Consequently, our survey presents both fundamental methodologies and their working, and gives practice-oriented recommendations on how to choose between different approaches for a problem at hand.

Carefully standardized facial images of 591 participants were taken in the laboratory, while controlling for self-presentation, facial expression, head orientation, and image properties. They were presented to human raters and a facial recognition algorithm: both humans (r=.21) and the algorithm (r=.22) could predict participants' scores on a political orientation scale (Cronbach's alpha=.94) decorrelated with age, gender, and ethnicity. These effects are on par with how well job interviews predict job success, or alcohol drives aggressiveness. Algorithm's predictive accuracy was even higher (r=.31) when it leveraged information on participants' age, gender, and ethnicity. Moreover, the associations between facial appearance and political orientation seem to generalize beyond our sample: The predictive model derived from standardized images (while controlling for age, gender, and ethnicity) could predict political orientation (r=.13) from naturalistic images of 3,401 politicians from the U.S., UK, and Canada. The analysis of facial features associated with political orientation revealed that conservatives tended to have larger lower faces. The predictability of political orientation from standardized images has critical implications for privacy, the regulation of facial recognition technology, and understanding the origins and consequences of political orientation.