Large Language Models (LLMs) benefit from training on ever larger amounts of textual data, but as a result, they increasingly incur the risk of leaking private information. The ability to selectively remove knowledge from LLMs is, therefore, a highly desirable capability. In this paper, we propose LUNAR, a novel unlearning methodology grounded in the Linear Representation Hypothesis. LUNAR operates by redirecting the representations of unlearned data to regions that trigger the model's inherent ability to express its inability to answer. LUNAR achieves state-of-the-art unlearning performance while significantly enhancing the controllability of the unlearned model during inference. Specifically, LUNAR achieves between 2.9x to 11.7x improvements on combined "unlearning efficacy" and "model utility" score ("Deviation Score") on the PISTOL dataset across various base models. We also demonstrate, through quantitative analysis and qualitative examples, LUNAR's superior controllability in generating coherent and contextually aware responses, mitigating undesired side effects of existing methods. Moreover, we demonstrate that LUNAR is robust against white-box adversarial attacks and versatile in handling real-world scenarios, such as processing sequential unlearning requests.

Anomaly detection (AD) is a critical machine learning task with diverse applications in web systems, including fraud detection, content moderation, and user behavior analysis. Despite its significance, AD in natural language processing (NLP) remains underexplored, limiting advancements in detecting anomalies in text data such as harmful content, phishing attempts, or spam reviews. In this paper, we introduce NLP-ADBench, the most comprehensive benchmark for NLP anomaly detection (NLP-AD), comprising eight curated datasets and evaluations of nineteen state-of-the-art algorithms. These include three end-to-end methods and sixteen two-step algorithms that apply traditional anomaly detection techniques to language embeddings generated by bert-base-uncased and OpenAI's text-embedding-3-large models. Our results reveal critical insights and future directions for NLP-AD. Notably, no single model excels across all datasets, highlighting the need for automated model selection. Moreover, two-step methods leveraging transformer-based embeddings consistently outperform specialized end-to-end approaches, with OpenAI embeddings demonstrating superior performance over BERT embeddings. By releasing NLP-ADBench at https://github.com/USC-FORTIS/NLP-ADBench, we provide a standardized framework for evaluating NLP-AD methods, fostering the development of innovative approaches. This work fills a crucial gap in the field and establishes a foundation for advancing NLP anomaly detection, particularly in the context of improving the safety and reliability of web-based systems.

Many well-established anomaly detection methods use the distance of a sample to those in its local neighbourhood: so-called `local outlier methods', such as LOF and DBSCAN. They are popular for their simple principles and strong performance on unstructured, feature-based data that is commonplace in many practical applications. However, they cannot learn to adapt for a particular set of data due to their lack of trainable parameters. In this paper, we begin by unifying local outlier methods by showing that they are particular cases of the more general message passing framework used in graph neural networks. This allows us to introduce learnability into local outlier methods, in the form of a neural network, for greater flexibility and expressivity: specifically, we propose LUNAR, a novel, graph neural network-based anomaly detection method. LUNAR learns to use information from the nearest neighbours of each node in a trainable way to find anomalies. We show that our method performs significantly better than existing local outlier methods, as well as state-of-the-art deep baselines. We also show that the performance of our method is much more robust to different settings of the local neighbourhood size.

With the advent of huges volumes of data produced in the form of fast streams, real-time machine learning has become a challenge of relevance emerging in a plethora of real-world applications. Processing such fast streams often demands high memory and processing resources. In addition, they can be affected by non-stationary phenomena (concept drift), by which learning methods have to detect changes in the distribution of streaming data, and adapt to these evolving conditions. A lack of efficient and scalable solutions is particularly noted in real-time scenarios where computing resources are severely constrained, as it occurs in networks of small, numerous, interconnected processing units (such as the so-called Smart Dust, Utility Fog, or Swarm Robotics paradigms). In this work we propose LUNAR, a streamified version of cellular automata devised to successfully meet the aforementioned requirements. It is able to act as a real incremental learner while adapting to drifting conditions. Extensive simulations with synthetic and real data will provide evidence of its competitive behavior in terms of classification performance when compared to long-established and successful online learning methods.

Have you ever wondered how most mammals are capable of fairly complex tasks, like walking, straight after being born? They haven't had time to experience the world yet so they've clearly not learnt how to perform the actions. Their brains must be pre-wired to enable them to walk, but if the brain structure relies on specific weights then an individual learning from its experiences could lose the ability to act shortly after birth, or never have the ability to begin with. Inspired by this, Adam Gaier and David Ha introduced the world to Weight Agnostic Neural Networks (WANN), an evolutionary strategy for developing neural networks which can perform a task independent of the weights of the connections. In this post, we'll briefly look into Weight Agnostic Neural Networks and use a code implementation to train our very own WANNs on the Lunar Lander gym environment.

Everything is more extreme on the moon. On top of temperatures that range from -300 F to 224 F, future astronauts and probes must deal with lighting conditions generously described as "harsh." To help, researchers at Ames Research Center in Silicon Valley created a lunar testbed, complete with craters, fluffy dust and solar simulator lights. The goal is to develop sensors that can "see" in such conditions to help probes and, eventually, humans navigate the surface safely. With no atmosphere to scatter and reflect lighting, "what you get on the Moon are dark shadows and very bright regions that are directly illuminated by the Sun -- the Italian painters in the Baroque period called it chiaroscuro," says NASA Ames computer scientist Uland Wong.