Publications

Creating secure and resilient applications with large language models (LLM) requires anticipating, adjusting to, and countering unforeseen threats. Red-teaming has emerged as a critical technique for identifying vulnerabilities in real-world LLM implementations. This paper presents a detailed threat model and provides a systematization of knowledge (SoK) of red-teaming attacks on LLMs. We develop a taxonomy of attacks based on the stages of the LLM development and deployment process and extract various insights from previous research. In addition, we compile methods for defense and practical red-teaming strategies for practitioners. By delineating prominent attack motifs and shedding light on various entry points, this paper provides a framework for improving the security and robustness of LLM-based systems.

Creating children's stories through text generation is a creative task that requires stories to be both entertaining and suitable for young audiences. However, since current story generation systems often rely on pre-trained language models fine-tuned with limited story data, they may not always prioritize child-friendliness. This can lead to the unintended generation of stories containing problematic elements such as violence, profanity, and biases. Regrettably, despite the significance of these concerns, there is a lack of clear guidelines and benchmark datasets for ensuring content safety for children. In this paper, we introduce a taxonomy specifically tailored to assess content safety in text, with a strong emphasis on children's well-being. We present PG-Story, a dataset that includes detailed annotations for both sentence-level and discourse-level safety. We demonstrate the potential of identifying unsafe content through self-diagnosis and employing controllable generation techniques during the decoding phase to minimize unsafe elements in generated stories.

Natural language often contains ambiguities that can lead to misinterpretation and miscommunication. While humans can handle ambiguities effectively by asking clarifying questions and/or relying on contextual cues and common-sense knowledge, resolving ambiguities can be notoriously hard for machines. In this work, we study ambiguities that arise in text-to-image generative models. We curate the Text-to-image Ambiguity Benchmark (TAB) dataset to study different types of ambiguities in text-to-image generative models. We then propose the Text-to-ImagE Disambiguation (TIED) framework to disambiguate the prompts given to the text-to-image generative models by soliciting clarifications from the end user. Through automatic and human evaluations, we show the effectiveness of our framework in generating more faithful images aligned with end user intention in the presence of ambiguities.

NLU models power several user facing experiences such as conversations agents and chat bots. Building NLU models typically consist of 3 stages: a) building or finetuning a pre-trained model b) distilling or fine-tuning the pre-trained model to build task specific models and, c) deploying the task-specific model to production. In this presentation, we will identify fairness considerations that can be incorporated in the aforementioned three stages in the life-cycle of NLU model building: (i) selection/building of a large scale language model, (ii) distillation/fine-tuning the large model into task specific model and, (iii) deployment of the task specific model. We will present select metrics that can be used to quantify fairness in NLU models and fairness enhancement techniques that can be deployed in each of these stages. Finally, we will share some recommendations to successfully implement fairness considerations when building an industrial scale NLU system.

In this work, we demonstrate that multilingual large-scale sequence-to-sequence (seq2seq) models, pre-trained on a mixture of denoising and Causal Language Modeling (CLM) tasks, are more efficient few-shot learners than decoder-only models on various tasks. In particular, we train a 20 billion parameter multilingual seq2seq model called Alexa Teacher Model (AlexaTM 20B) and show that it achieves state-of-the-art (SOTA) performance on 1-shot summarization tasks, outperforming a much larger 540B PaLM decoder model. AlexaTM 20B also achieves SOTA in 1-shot machine translation, especially for low-resource languages, across almost all language pairs supported by the model (Arabic, English, French, German, Hindi, Italian, Japanese, Marathi, Portuguese, Spanish, Tamil, and Telugu) on Flores-101 dataset. We also show in zero-shot setting, AlexaTM 20B outperforms GPT3 (175B) on SuperGLUE and SQuADv2 datasets and provides SOTA performance on multilingual tasks such as XNLI, XCOPA, Paws-X, and XWinograd. Overall, our results present a compelling case for seq2seq models as a powerful alternative to decoder-only models for Large-scale Language Model (LLM) training.

With the rapid growth in language processing applications, fairness has emerged as an important consideration in data-driven solutions. Although various fairness definitions have been explored in the recent literature, there is lack of consensus on which metrics most accurately reflect the fairness of a system. In this work, we propose a new formulation -- accumulated prediction sensitivity, which measures fairness in machine learning models based on the model's prediction sensitivity to perturbations in input features. The metric attempts to quantify the extent to which a single prediction depends on a protected attribute, where the protected attribute encodes the membership status of an individual in a protected group. We show that the metric can be theoretically linked with a specific notion of group fairness (statistical parity) and individual fairness. It also correlates well with humans' perception of fairness. We conduct experiments on two text classification datasets -- Jigsaw Toxicity, and Bias in Bios, and evaluate the correlations between metrics and manual annotations on whether the model produced a fair outcome. We observe that the proposed fairness metric based on prediction sensitivity is statistically significantly more correlated with human annotation than the existing counterfactual fairness metric.

Language models excel at generating coherent text, and model compression techniques such as knowledge distillation have enabled their use in resource-constrained settings. However, these models can be biased in multiple ways, including the unfounded association of male and female genders with gender-neutral professions. Therefore, knowledge distillation without any fairness constraints may preserve or exaggerate the teacher model's biases onto the distilled model. To this end, we present a novel approach to mitigate gender disparity in text generation by learning a fair model during knowledge distillation. We propose two modifications to the base knowledge distillation based on counterfactual role reversal---modifying teacher probabilities and augmenting the training set. We evaluate gender polarity across professions in open-ended text generated from the resulting distilled and finetuned GPT-2 models and demonstrate a substantial reduction in gender disparity with only a minor compromise in utility. Finally, we observe that language models that reduce gender polarity in language generation do not improve embedding fairness or downstream classification fairness.

Recently, there is a surge of interest in using point processes to model continuous-time user activities. This framework has resulted in novel models and improved performance in diverse applications. However, most previous works focus on the 'open loop' setting where learned models are used for predictive tasks. Typically, we are interested in the 'closed loop' setting where a policy needs to be learned to incorporate user feedbacks and guide user activities to desirable states. Although point processes have good predictive performance, it is not clear how to use them for the challenging closed loop activity guiding task. In this paper, we propose a framework to reformulate point processes into stochastic differential equations, which allows us to extend methods from stochastic optimal control to address the activity guiding problem. We also design an efficient algorithm, and show that our method guides user activities to desired states more effectively than state-of-arts.

Large volumes of networked streaming event data are becoming increasingly available in a wide variety of applications such as social network analysis, Internet traffic monitoring, and health care analytics. Streaming event data are discrete observations occurring in continuous time, and the precise time interval between two events carries substantial information about the dynamics of the underlying systems. How does one promptly detect changes in these dynamic systems using these streaming event data? In this paper, we propose a novel change-point detection framework for multidimensional event data over networks. We cast the problem into a sequential hypothesis test, and we derive the likelihood ratios for point processes, which are computed efficiently via an expectation-maximization (EM) like algorithm that is parameter free and can be computed in a distributed manner. We derive a highly accurate theoretical characterization of the falsealarm rate, and we show that the method can provide weak signal detection by aggregating local statistics over time and networks. Finally, we demonstrate the good performance of our algorithm on numerical examples and real-world datasets from Twitter and Memetracker.