This research addresses the challenge of integrating forecasting and optimization in energy management systems, focusing on the impacts of switching costs, forecast accuracy, and stability. It proposes a novel framework for analyzing online optimization problems with switching costs and enabled by deterministic and probabilistic forecasts. Through empirical evaluation and theoretical analysis, the research reveals the balance between forecast accuracy, stability, and switching costs in shaping policy performance. Conducted in the context of battery scheduling within energy management applications, it introduces a metric for evaluating probabilistic forecast stability and examines the effects of forecast accuracy and stability on optimization outcomes using the real-world case of the Citylearn 2022 competition. Findings indicate that switching costs significantly influence the trade-off between forecast accuracy and stability, highlighting the importance of integrated systems that enable collaboration between forecasting and operational units for improved decision-making. The study shows that committing to a policy for longer periods can be advantageous over frequent updates. Results also show a correlation between forecast stability and policy performance, suggesting that stable forecasts can mitigate switching costs. The proposed framework provides valuable insights for energy sector decision-makers and forecast practitioners when designing the operation of an energy management system.

Large language models (LLMs) are prone to hallucinations, which sparked a widespread effort to detect and prevent them. Recent work attempts to mitigate hallucinations by intervening in the model's generation, typically computing representative vectors of hallucinations vs. grounded generations, for steering the model's hidden states away from a hallucinatory state. However, common studies employ different setups and do not properly separate different possible causes of hallucinations, making interventions misguided. In this work, we introduce a method for categorizing examples based on the model's prior knowledge, named WACK. We construct WACK benchmarks that support interventions in two settings: open-book and closed-book question answering. Using the benchmarks, we perform an extensive investigation of the effect of different choices for intervention, such as the intervened components, and how often and how strongly to intervene. We find that intervention success varies depending on the component, with the attention blocks performing well and the residual stream proving detrimental to language modeling capabilities. We also show that interventions can benefit from representative vectors collected before, rather than after, a hallucination occurs. Finally, we introduce a new dynamic intervention, which intervenes only if needed, and thus is more robust than standard static interventions.

Argumentation is a formalism allowing to reason with contradictory information by modeling arguments and their interactions. There are now an increasing number of gradual semantics and impact measures that have emerged to facilitate the interpretation of their outcomes. An impact measure assesses, for each argument, the impact of other arguments on its score. In this paper, we refine an existing impact measure from Delobelle and Villata and introduce a new impact measure rooted in Shapley values. We introduce several principles to evaluate those two impact measures w.r.t. some well-known gradual semantics. This comprehensive analysis provides deeper insights into their functionality and desirability.

Decentralized planning is a key element of cooperative multi-agent systems for information gathering tasks. However, despite the high frequency of agent failures in realistic large deployment scenarios, current approaches perform poorly in the presence of failures, by not converging at all, and/or by making very inefficient use of resources (e.g. energy). In this work, we propose Attritable MCTS (A-MCTS), a decentralized MCTS algorithm capable of timely and efficient adaptation to changes in the set of active agents. It is based on the use of a global reward function for the estimation of each agent's local contribution, and regret matching for coordination. We evaluate its effectiveness in realistic data-harvesting problems under different scenarios. We show both theoretically and experimentally that A-MCTS enables efficient adaptation even under high failure rates. Results suggest that, in the presence of frequent failures, our solution improves substantially over the best existing approaches in terms of global utility and scalability.

Cryptic crossword clues are challenging cognitive tasks, for which new test sets are released on a daily basis by multiple international newspapers. Each cryptic clue contains both the definition of the answer to be placed in the crossword grid (in common with regular crosswords), and `wordplay' that proves that the answer is correct (i.e. a human solver can be confident that an answer is correct without needing crossing words to confirm it). Using an existing cryptic wordplay proving framework (operating on Python proofs created by an LLM), we show that it is possible to distinguish between correct answers and almost-correct ones based upon whether the wordplay `works'.

Retrieval Augmented Generation (RAG) techniques aim to mitigate hallucinations in Large Language Models (LLMs). However, LLMs can still produce information that is unsupported or contradictory to the retrieved contexts. We introduce LYNX, a SOTA hallucination detection LLM that is capable of advanced reasoning on challenging real-world hallucination scenarios. To evaluate LYNX, we present HaluBench, a comprehensive hallucination evaluation benchmark, consisting of 15k samples sourced from various real-world domains. Our experiment results show that LYNX outperforms GPT-4o, Claude-3-Sonnet, and closed and open-source LLM-as-a-judge models on HaluBench. We release LYNX, HaluBench and our evaluation code for public access.

We introduce Modelizer - a novel framework that, given a black-box program, learns a _model from its input/output behavior_ using _neural machine translation_. The resulting model _mocks_ the original program: Given an input, the model predicts the output that would have been produced by the program. However, the model is also _reversible_ - that is, the model can predict the input that would have produced a given output. Finally, the model is _differentiable_ and can be efficiently restricted to predict only a certain aspect of the program behavior. Modelizer uses _grammars_ to synthesize inputs and to parse the resulting outputs, allowing it to learn sequence-to-sequence associations between token streams. Other than input and output grammars, Modelizer only requires the ability to execute the program. The resulting models are _small_, requiring fewer than 6.3 million parameters for languages such as Markdown or HTML; and they are _accurate_, achieving up to 95.4% accuracy and a BLEU score of 0.98 with standard error 0.04 in mocking real-world applications. We foresee several _applications_ of these models, especially as the output of the program can be any aspect of program behavior. Besides mocking and predicting program behavior, the model can also synthesize inputs that are likely to produce a particular behavior, such as failures or coverage.

The main goal of diversity optimization is to find a diverse set of solutions which satisfy some lower bound on their fitness. Evolutionary algorithms (EAs) are often used for such tasks, since they are naturally designed to optimize populations of solutions. This approach to diversity optimization, called EDO, has been previously studied from theoretical perspective, but most studies considered only EAs with a trivial offspring population such as the $(\mu + 1)$ EA. In this paper we give an example instance of a $k$-vertex cover problem, which highlights a critical difference of the diversity optimization from the regular single-objective optimization, namely that there might be a locally optimal population from which we can escape only by replacing at least two individuals at once, which the $(\mu + 1)$ algorithms cannot do. We also show that the $(\mu + \lambda)$ EA with $\lambda \ge \mu$ can effectively find a diverse population on $k$-vertex cover, if using a mutation operator inspired by Branson and Sutton (TCS 2023). To avoid the problem of subset selection which arises in the $(\mu + \lambda)$ EA when it optimizes diversity, we also propose the $(1_\mu + 1_\mu)$ EA$_D$, which is an analogue of the $(1 + 1)$ EA for populations, and which is also efficient at optimizing diversity on the $k$-vertex cover problem.

The challenge of finding content that aligns with users' interests within this abundance has become increasingly important. Recommender systems play a crucial role in addressing this issue, as they have the potential to provide precise recommendations that enhance user experience and save time in commercial applications [1]. These systems predict user ratings for specific items by employing data mining techniques and related predictive algorithms to make highly relevant predictions. By analyzing user historical behavior, preferences, and item characteristics, recommender systems effectively solve the information filtering problem by automatically matching items that may be of interest to users. Traditional recommender systems primarily consist of collaborative filtering [2], content-based recommendations [3], and hybrid recommendation methods, among which collaborative filtering is one of the earliest and most widely used techniques for recommending products or items based on past purchasing history.

Currently, image-text-driven multi-modal deep learning models have demonstrated their outstanding potential in many fields. In practice, tasks centered around facial images have broad application prospects. This paper presents \textbf{FaceCaption-15M}, a large-scale, diverse, and high-quality dataset of facial images accompanied by their natural language descriptions (facial image-to-text). This dataset aims to facilitate a study on face-centered tasks. FaceCaption-15M comprises over 15 million pairs of facial images and their corresponding natural language descriptions of facial features, making it the largest facial image-caption dataset to date. We conducted a comprehensive analysis of image quality, text naturalness, text complexity, and text-image relevance to demonstrate the superiority of FaceCaption-15M. To validate the effectiveness of FaceCaption-15M, we first trained a facial language-image pre-training model (FLIP, similar to CLIP) to align facial image with its corresponding captions in feature space. Subsequently, using both image and text encoders and fine-tuning only the linear layer, our FLIP-based models achieved state-of-the-art results on two challenging face-centered tasks. The purpose is to promote research in the field of face-related tasks through the availability of the proposed FaceCaption-15M dataset. All data, codes, and models are publicly available. https://huggingface.co/datasets/OpenFace-CQUPT/FaceCaption-15M