Search

Multilingual LLMs often have knowledge disparities across languages, with larger gaps in under-resourced languages. Teaching LLMs to abstain in the face of knowledge gaps is thus a promising strategy to mitigate hallucinations in multilingual settings. However, previous studies on LLM abstention primarily focus on English; we find that directly applying existing solutions beyond English results in up to 20.5% performance gaps between high and low-resource languages, potentially due to LLMs' drop in calibration and reasoning beyond a few resource-rich languages. To this end, we propose strategies to enhance LLM abstention by learning from multilingual feedback, where LLMs self-reflect on proposed answers in one language by generating multiple feedback items in related languages: we show that this helps identifying the knowledge gaps across diverse languages, cultures, and communities. Extensive experiments demonstrate that our multilingual feedback approach outperforms various strong baselines, achieving up to 9.2% improvement for low-resource languages across three black-box and open models on three datasets, featuring open-book, closed-book, and commonsense QA. Further analysis reveals that multilingual feedback is both an effective and a more equitable abstain strategy to serve diverse language speakers, and cultural factors have great impact on language selection and LLM abstention behavior, highlighting future directions for multilingual and multi-cultural reliable language modeling.

Despite efforts to expand the knowledge of large language models (LLMs), knowledge gaps -- missing or outdated information in LLMs -- might always persist given the evolving nature of knowledge. In this work, we study approaches to identify LLM knowledge gaps and abstain from answering questions when knowledge gaps are present. We first adapt existing approaches to model calibration or adaptation through fine-tuning/prompting and analyze their ability to abstain from generating low-confidence outputs. Motivated by their failures in self-reflection and over-reliance on held-out sets, we propose two novel approaches that are based on model collaboration, i.e., LLMs probing other LLMs for knowledge gaps, either cooperatively or competitively. Extensive experiments with three LLMs on four QA tasks featuring diverse knowledge domains demonstrate that both cooperative and competitive approaches to unveiling LLM knowledge gaps achieve up to 19.3% improvements on abstain accuracy against the strongest baseline. Further analysis reveals that our proposed mechanisms could help identify failure cases in retrieval augmentation and pinpoint knowledge gaps in multi-hop reasoning., Comment: ACL 2024

This paper presents a novel application of graph neural networks for modeling and estimating network heterogeneity. Network heterogeneity is characterized by variations in unit's decisions or outcomes that depend not only on its own attributes but also on the conditions of its surrounding neighborhood. We delineate the convergence rate of the graph neural networks estimator, as well as its applicability in semiparametric causal inference with heterogeneous treatment effects. The finite-sample performance of our estimator is evaluated through Monte Carlo simulations. In an empirical setting related to microfinance program participation, we apply the new estimator to examine the average treatment effects and outcomes of counterfactual policies, and to propose an enhanced strategy for selecting the initial recipients of program information in social networks.

Large language models (LLMs) often encounter knowledge conflicts, scenarios where discrepancy arises between the internal parametric knowledge of LLMs and non-parametric information provided in the prompt context. In this work we ask what are the desiderata for LLMs when a knowledge conflict arises and whether existing LLMs fulfill them. We posit that LLMs should 1) identify knowledge conflicts, 2) pinpoint conflicting information segments, and 3) provide distinct answers or viewpoints in conflicting scenarios. To this end, we introduce KNOWLEDGE CONFLICT, an evaluation framework for simulating contextual knowledge conflicts and quantitatively evaluating to what extent LLMs achieve these goals. KNOWLEDGE CONFLICT includes diverse and complex situations of knowledge conflict, knowledge from diverse entities and domains, two synthetic conflict creation methods, and settings with progressively increasing difficulty to reflect realistic knowledge conflicts. Extensive experiments with the KNOWLEDGE CONFLICT framework reveal that while LLMs perform well in identifying the existence of knowledge conflicts, they struggle to determine the specific conflicting knowledge and produce a response with distinct answers amidst conflicting information. To address these challenges, we propose new instruction-based approaches that augment LLMs to better achieve the three goals. Further analysis shows that abilities to tackle knowledge conflicts are greatly impacted by factors such as knowledge domain and prompt text, while generating robust responses to knowledge conflict scenarios remains an open research question.

In a pre-registered randomized experiment, we found that, relative to a reverse-chronological baseline, Twitter's engagement-based ranking algorithm amplifies emotionally charged, out-group hostile content that users say makes them feel worse about their political out-group. Furthermore, we find that users do not prefer the political tweets selected by the algorithm, suggesting that the engagement-based algorithm underperforms in satisfying users' stated preferences. Finally, we explore the implications of an alternative approach that ranks content based on users' stated preferences and find a reduction in angry, partisan, and out-group hostile content but also a potential reinforcement of echo chambers. The evidence underscores the necessity for a more nuanced approach to content ranking that balances engagement, users' stated preferences, and sociopolitical outcomes.

To improve the spatial resolution of power Doppler (PD) imaging, we explored null subtraction imaging (NSI) as an alternative beamforming technique to delay-and-sum (DAS). NSI is a nonlinear beamforming approach that uses three different apodizations on receive and incoherently sums the beamformed envelopes. NSI uses a null in the beam pattern to improve the lateral resolution, which we apply here for improving PD spatial resolution both with and without contrast microbubbles. In this study, we used NSI with three types of singular value decomposition (SVD)-based clutter filters and noise equalization to generate high-resolution PD images. An element sensitivity correction scheme was also proposed as a crucial component of NSI-based PD imaging. First, a microbubble trace experiment was performed to evaluate the resolution improvement of NSI-based PD over traditional DAS-based PD. Then, both contrast-enhanced and contrast free ultrasound PD images were generated from the scan of a rat brain. The cross-sectional profile of the microbubble traces and microvessels were plotted. FWHM was also estimated to provide a quantitative metric. Furthermore, iso-frequency curves were calculated to provide a resolution evaluation metric over the global field of view. Up to six-fold resolution improvement was demonstrated by the FWHM estimate and four-fold resolution improvement was demonstrated by the iso-frequency curve from the NSI-based PD microvessel images compared to microvessel images generated by traditional DAS-based beamforming. A resolvability of 39 um was measured from the NSI-based PD microvessel image. The computational cost of NSI-based PD was only increased by 40 percent over the DAS-based PD.

Semantic segmentation is important for scene understanding. To address the scenes of adverse illumination conditions of natural images, thermal infrared (TIR) images are introduced. Most existing RGB-T semantic segmentation methods follow three cross-modal fusion paradigms, i.e. encoder fusion, decoder fusion, and feature fusion. Some methods, unfortunately, ignore the properties of RGB and TIR features or the properties of features at different levels. In this paper, we propose a novel feature fusion-based network for RGB-T semantic segmentation, named \emph{LASNet}, which follows three steps of location, activation, and sharpening. The highlight of LASNet is that we fully consider the characteristics of cross-modal features at different levels, and accordingly propose three specific modules for better segmentation. Concretely, we propose a Collaborative Location Module (CLM) for high-level semantic features, aiming to locate all potential objects. We propose a Complementary Activation Module for middle-level features, aiming to activate exact regions of different objects. We propose an Edge Sharpening Module (ESM) for low-level texture features, aiming to sharpen the edges of objects. Furthermore, in the training phase, we attach a location supervision and an edge supervision after CLM and ESM, respectively, and impose two semantic supervisions in the decoder part to facilitate network convergence. Experimental results on two public datasets demonstrate that the superiority of our LASNet over relevant state-of-the-art methods. The code and results of our method are available at https://github.com/MathLee/LASNet., Comment: 12 pages, 7 figures, Accepted by IEEE Transactions on Circuits and Systems for Video Technology 2022