AI-generated text detection has attracted increasing attention as powerful language models approach human-level generation. Limited work is devoted to detecting (partially) AI-paraphrased texts. However, AI paraphrasing is commonly employed in various application scenarios for text refinement and diversity. To this end, we propose a novel detection framework, paraphrased text span detection (PTD), aiming to identify paraphrased text spans within a text. Different from text-level detection, PTD takes in the full text and assigns each of the sentences with a score indicating the paraphrasing degree. We construct a dedicated dataset, PASTED, for paraphrased text span detection. Both in-distribution and out-of-distribution results demonstrate the effectiveness of PTD models in identifying AI-paraphrased text spans. Statistical and model analysis explains the crucial role of the surrounding context of the paraphrased text spans. Extensive experiments show that PTD models can generalize to versatile paraphrasing prompts and multiple paraphrased text spans. We release our resources at https://github.com/Linzwcs/PASTED.

Current LLMs are generally aligned to follow safety requirements and tend to refuse toxic prompts. However, LLMs can fail to refuse toxic prompts or be overcautious and refuse benign examples. In addition, state-of-the-art toxicity detectors have low TPRs at low FPR, incurring high costs in real-world applications where toxic examples are rare. In this paper, we explore Moderation Using LLM Introspection (MULI), which detects toxic prompts using the information extracted directly from LLMs themselves. We found significant gaps between benign and toxic prompts in the distribution of alternative refusal responses and in the distribution of the first response token's logits. These gaps can be used to detect toxicities: We show that a toy model based on the logits of specific starting tokens gets reliable performance, while requiring no training or additional computational cost. We build a more robust detector using a sparse logistic regression model on the first response token logits, which greatly exceeds SOTA detectors under multiple metrics.

Creating fair AI systems is a complex problem that involves the assessment of context-dependent bias concerns. Existing research and programming libraries express specific concerns as measures of bias that they aim to constrain or mitigate. In practice, one should explore a wide variety of (sometimes incompatible) measures before deciding which ones warrant corrective action, but their narrow scope means that most new situations can only be examined after devising new measures. In this work, we present a mathematical framework that distils literature measures of bias into building blocks, hereby facilitating new combinations to cover a wide range of fairness concerns, such as classification or recommendation differences across multiple multi-value sensitive attributes (e.g., many genders and races, and their intersections). We show how this framework generalizes existing concepts and present frequently used blocks. We provide an open-source implementation of our framework as a Python library, called FairBench, that facilitates systematic and extensible exploration of potential bias concerns.

Recent end-to-end approaches have shown promise in extending large language models (LLMs) to speech inputs, but face limitations in directly assessing and optimizing alignment quality and fail to achieve fine-grained alignment due to speech-text length mismatch. We introduce BLSP-KD, a novel approach for Bootstrapping Language-Speech Pretraining via Knowledge Distillation, which addresses these limitations through two key techniques. First, it optimizes speech-text alignment by minimizing the divergence between the LLM's next-token prediction distributions for speech and text inputs using knowledge distillation. Second, it employs a continuous-integrate-andfire strategy to segment speech into tokens that correspond one-to-one with text tokens, enabling fine-grained alignment. We also introduce Partial LoRA (PLoRA), a new adaptation method supporting LLM finetuning for speech inputs under knowledge distillation. Quantitative evaluation shows that BLSP-KD outperforms previous end-to-end baselines and cascaded systems with comparable scale of parameters, facilitating general instruction-following capabilities for LLMs with speech inputs. This approach provides new possibilities for extending LLMs to spoken language interactions.

Deferring systems extend supervised Machine Learning (ML) models with the possibility to defer predictions to human experts. However, evaluating the impact of a deferring strategy on system accuracy is still an overlooked area. This paper fills this gap by evaluating deferring systems through a causal lens. We link the potential outcomes framework for causal inference with deferring systems. This allows us to identify the causal impact of the deferring strategy on predictive accuracy. We distinguish two scenarios. In the first one, we can access both the human and the ML model predictions for the deferred instances. In such a case, we can identify the individual causal effects for deferred instances and aggregates of them. In the second scenario, only human predictions are available for the deferred instances. In this case, we can resort to regression discontinuity design to estimate a local causal effect. We empirically evaluate our approach on synthetic and real datasets for seven deferring systems from the literature.

In applied statistics and machine learning, the "gold standards" used for training are often biased and almost always noisy. Dawid and Skene's justifiably popular crowdsourcing model adjusts for rater (coder, annotator) sensitivity and specificity, but fails to capture distributional properties of rating data gathered for training, which in turn biases training. In this study, we introduce a general purpose measurement-error model with which we can infer consensus categories by adding item-level effects for difficulty, discriminativeness, and guessability. We further show how to constrain the bimodal posterior of these models to avoid (or if necessary, allow) adversarial raters. We validate our model's goodness of fit with posterior predictive checks, the Bayesian analogue of $\chi^2$ tests. Dawid and Skene's model is rejected by goodness of fit tests, whereas our new model, which adjusts for item heterogeneity, is not rejected. We illustrate our new model with two well-studied data sets, binary rating data for caries in dental X-rays and implication in natural language.

In this paper, we study a class of non-smooth non-convex problems in the form of $\min_{x}[\max_{y\in Y}\phi(x, y) - \max_{z\in Z}\psi(x, z)]$, where both $\Phi(x) = \max_{y\in Y}\phi(x, y)$ and $\Psi(x)=\max_{z\in Z}\psi(x, z)$ are weakly convex functions, and $\phi(x, y), \psi(x, z)$ are strongly concave functions in terms of $y$ and $z$, respectively. It covers two families of problems that have been studied but are missing single-loop stochastic algorithms, i.e., difference of weakly convex functions and weakly convex strongly-concave min-max problems. We propose a stochastic Moreau envelope approximate gradient method dubbed SMAG, the first single-loop algorithm for solving these problems, and provide a state-of-the-art non-asymptotic convergence rate. The key idea of the design is to compute an approximate gradient of the Moreau envelopes of $\Phi, \Psi$ using only one step of stochastic gradient update of the primal and dual variables. Empirically, we conduct experiments on positive-unlabeled (PU) learning and partial area under ROC curve (pAUC) optimization with an adversarial fairness regularizer to validate the effectiveness of our proposed algorithms.

Recent works have extended notions of feature importance to \emph{semantic concepts} that are inherently interpretable to the users interacting with a black-box predictive model. Yet, precise statistical guarantees, such as false positive rate control, are needed to communicate findings transparently and to avoid unintended consequences in real-world scenarios. In this paper, we formalize the global (i.e., over a population) and local (i.e., for a sample) statistical importance of semantic concepts for the predictions of opaque models, by means of conditional independence, which allows for rigorous testing. We use recent ideas of sequential kernelized testing (SKIT) to induce a rank of importance across concepts, and showcase the effectiveness and flexibility of our framework on synthetic datasets as well as on image classification tasks using vision-language models such as CLIP.

Physical neural networks (PNNs) are emerging paradigms for neural network acceleration due to their high-bandwidth, in-propagation analogue processing. Despite the advantages of PNN for inference, training remains a challenge. The imperfect information of the physical transformation means the failure of conventional gradient-based updates from backpropagation (BP). Here, we present the asymmetrical training (AT) method, which treats the PNN structure as a grey box. AT performs training while only knowing the last layer output and neuron topological connectivity of a deep neural network structure, not requiring information about the physical control-transformation mapping. We experimentally demonstrated the AT method on deep grey-box PNNs implemented by uncalibrated photonic integrated circuits (PICs), improving the classification accuracy of Iris flower and modified MNIST hand-written digits from random guessing to near theoretical maximum. We also showcased the consistently enhanced performance of AT over BP for different datasets, including MNIST, fashion-MNIST, and Kuzushiji-MNIST. The AT method demonstrated successful training with minimal hardware overhead and reduced computational overhead, serving as a robust light-weight training alternative to fully explore the advantages of physical computation.

Entity alignment (EA) aims to merge two knowledge graphs (KGs) by identifying equivalent entity pairs. While existing methods heavily rely on human-generated labels, it is prohibitively expensive to incorporate cross-domain experts for annotation in real-world scenarios. The advent of Large Language Models (LLMs) presents new avenues for automating EA with annotations, inspired by their comprehensive capability to process semantic information. However, it is nontrivial to directly apply LLMs for EA since the annotation space in real-world KGs is large. LLMs could also generate noisy labels that may mislead the alignment. To this end, we propose a unified framework, LLM4EA, to effectively leverage LLMs for EA. Specifically, we design a novel active learning policy to significantly reduce the annotation space by prioritizing the most valuable entities based on the entire inter-KG and intra-KG structure. Moreover, we introduce an unsupervised label refiner to continuously enhance label accuracy through in-depth probabilistic reasoning. We iteratively optimize the policy based on the feedback from a base EA model. Extensive experiments demonstrate the advantages of LLM4EA on four benchmark datasets in terms of effectiveness, robustness, and efficiency. Codes are available via https://github.com/chensyCN/llm4ea_official.