We present MiniCalc, a web app for teaching first-order logic, based on a so-called minimal sequent calculus. We explain the sequent calculus in Section 2. More than 100 computer science students have used versions of MiniCalc in a course on automated reasoning in the period 2021-2024. The web app MiniCalc 1.0 has not yet been announced, but it is available here: https://proof.compute.dtu.dk/MiniCalc.zip Installation is easy: Just unpack MiniCalc.zip in a new directory and open index.html in a browser. MiniCalc displays the proof editor to the left and the result about the default example proof to the right. We explain the default example proof in Section 3. The files in the above zip are from 12 February 2024 and we are not aware of bugs as of 1 December 2024.

We propose a method for training language models in an interactive setting inspired by child language acquisition. In our setting, a speaker attempts to communicate some information to a listener in a single-turn dialogue and receives a reward if communicative success is achieved. Unlike earlier related work using image--caption data for interactive reference games, we operationalize communicative success in a more abstract language-only question--answering setting. First, we present a feasibility study demonstrating that our reward provides an indirect signal about grammaticality. Second, we conduct experiments using reinforcement learning to fine-tune language models. We observe that cognitively plausible constraints on the communication channel lead to interpretable changes in speaker behavior. However, we do not yet see improvements on linguistic evaluations from our training regime. We outline potential modifications to the task design and training configuration that could better position future work to use our methodology to observe the benefits of interaction on language learning in computational cognitive models.

Evaluating Retrieval-Augmented Generation (RAG) in large language models (LLMs) is challenging because benchmarks can quickly become stale. Questions initially requiring retrieval may become answerable from pretraining knowledge as newer models incorporate more recent information during pretraining, making it difficult to distinguish evidence-based reasoning from recall. We introduce NeoQA (News Events for Out-of-training Question Answering), a benchmark designed to address this issue. To construct NeoQA, we generated timelines and knowledge bases of fictional news events and entities along with news articles and Q\&A pairs to prevent LLMs from leveraging pretraining knowledge, ensuring that no prior evidence exists in their training data. We propose our dataset as a new platform for evaluating evidence-based question answering, as it requires LLMs to generate responses exclusively from retrieved evidence and only when sufficient evidence is available. NeoQA enables controlled evaluation across various evidence scenarios, including cases with missing or misleading details. Our findings indicate that LLMs struggle to distinguish subtle mismatches between questions and evidence, and suffer from short-cut reasoning when key information required to answer a question is missing from the evidence, underscoring key limitations in evidence-based reasoning.

--Recently, spiking neural networks (SNNs), deployed on neuromorphic chips, provide highly efficient solutions on edge devices in different scenarios. However, their ability to adapt to distribution shifts after deployment has become a crucial challenge. Online test-time adaptation (OTT A) offers a promising solution by enabling models to dynamically adjust to new data distributions without requiring source data or labeled target samples. Nevertheless, existing OTT A methods are largely designed for traditional artificial neural networks and are not well-suited for SNNs. T o address this gap, we propose a low-power, neuromorphic chip-friendly online test-time adaptation framework, aiming to enhance model generalization under distribution shifts. The proposed approach is called Threshold Modulation (TM), which dynamically adjusts the firing threshold through neuronal dynamics-inspired normalization, being more compatible with neuromorphic hardware. Experimental results on benchmark datasets demonstrate the effectiveness of this method in improving the robustness of SNNs against distribution shifts while maintaining low computational cost. The proposed method offers a practical solution for online test-time adaptation of SNNs, providing inspiration for the design of future neuromorphic chips. In recent years, with the rapid development of high-performance hardware and training algorithms, modern deep artifical neural networks (ANNs) can have billions, or even hundreds of billions, of parameters, requiring large-scale computational resource for training and inference.

Particle robots are novel biologically-inspired robotic systems where locomotion can be achieved collectively and robustly, but not independently. While its control is currently limited to a hand-crafted policy for basic locomotion tasks, such a multi-robot system could be potentially controlled via Deep Reinforcement Learning (DRL) for different tasks more efficiently. However, the particle robot system presents a new set of challenges for DRL differing from existing swarm robotics systems: the low degrees of freedom of each robot and the increased necessity of coordination between robots. We present a 2D particle robot simulator using the OpenAI Gym interface and Pymunk as the physics engine, and introduce new tasks and challenges to research the underexplored applications of DRL in the particle robot system. Moreover, we use Stable-baselines3 to provide a set of benchmarks for the tasks. Current baseline DRL algorithms show signs of achieving the tasks but are yet unable to reach the performance of the hand-crafted policy. Further development of DRL algorithms is necessary in order to accomplish the proposed tasks.

In performative Reinforcement Learning (RL), an agent faces a policy-dependent environment: the reward and transition functions depend on the agent's policy. Prior work on performative RL has studied the convergence of repeated retraining approaches to a performatively stable policy. In the finite sample regime, these approaches repeatedly solve for a saddle point of a convex-concave objective, which estimates the Lagrangian of a regularized version of the reinforcement learning problem. In this paper, we aim to extend such repeated retraining approaches, enabling them to operate under corrupted data. More specifically, we consider Huber's $ε$-contamination model, where an $ε$ fraction of data points is corrupted by arbitrary adversarial noise. We propose a repeated retraining approach based on convex-concave optimization under corrupted gradients and a novel problem-specific robust mean estimator for the gradients. We prove that our approach exhibits last-iterate convergence to an approximately stable policy, with the approximation error linear in $\sqrtε$. We experimentally demonstrate the importance of accounting for corruption in performative RL.

While the use of MLdriven systems can enhance efficiency, it can also drive the humans who are subject to algorithmic decisions to adjust their behavior accordingly. Examples include Uber drivers coordinating their behavior in response to its surge pricing algorithm [Möhlmann and Zalmanson, 2017], applicants selecting keywords and formatting to pass automated resume screening [Forbes, 2022], and Facebook users adjusting their posting and content interaction choices in response to the platforms' curation algorithms [Eslami et al., 2016]. These can be viewed as strategic responses by rational human subjects in these systems, motivating a game-theoretical analysis of learning algorithms with human in the loop. Earlier works on the study of strategic humans facing ML systems largely focused on scenarios where users can strategically alter only their observable data (e.g., students cheating to obtain better test scores, job applicants making formatting or wording changes to their CV, or loan applicants opening several new accounts to increase their credit scores) to receive a favorable decision (e.g., be accepted to a school, job opening, or loan); see, e.g., [Hu et al., 2019, Milli et al., 2019]. This strategic behavior is referred to as strategic manipulation, where agents change their features without changing their true qualification states. This can be interpreted as cheating the machine learning algorithm: such agents may appear to be more qualified, without being truly suitable for a favorable outcome.

Trust is a fundamental component of human-agent interaction. With the increasing presence of artificial agents in daily life, it is essential to understand how people perceive and trust these agents. One of the key challenges affecting this perception is the Uncanny Valley Effect (UVE), where increasingly human-like artificial beings can be perceived as eerie or repelling. Despite growing interest in trust and the UVE, existing research varies widely in terms of how these concepts are defined and operationalized. This inconsistency raises important questions about how and under what conditions the UVE influences trust in agents. A systematic understanding of their relationship is currently lacking. This review aims to examine the impact of the UVE on human trust in agents and to identify methodological patterns, limitations, and gaps in the existing empirical literature. Following PRISMA guidelines, a systematic search identified 53 empirical studies that investigated both UVE-related constructs and trust or trust-related outcomes. Studies were analyzed based on a structured set of categories, including types of agents and interactions, methodological and measurement approaches, and key findings. The results of our systematic review reveal that most studies rely on static images or hypothetical scenarios with limited real-time interaction, and the majority use subjective trust measures. This review offers a novel framework for classifying trust measurement approaches with regard to the best-practice criteria for empirically investigating the UVE. As the first systematic attempt to map the intersection of UVE and trust, this review contributes to a deeper understanding of their interplay and offers a foundation for future research. Keywords: the uncanny valley effect, trust, human-likeness, affinity response, human-agent interaction

As frontier AI systems advance toward transformative capabilities, we need a parallel transformation in how we measure and evaluate these systems to ensure safety and inform governance. While benchmarks have been the primary method for estimating model capabilities, they often fail to establish true upper bounds or predict deployment behavior. This literature review consolidates the rapidly evolving field of AI safety evaluations, proposing a systematic taxonomy around three dimensions: what properties we measure, how we measure them, and how these measurements integrate into frameworks. We show how evaluations go beyond benchmarks by measuring what models can do when pushed to the limit (capabilities), the behavioral tendencies exhibited by default (propensities), and whether our safety measures remain effective even when faced with subversive adversarial AI (control). These properties are measured through behavioral techniques like scaffolding, red teaming and supervised fine-tuning, alongside internal techniques such as representation analysis and mechanistic interpretability. We provide deeper explanations of some safety-critical capabilities like cybersecurity exploitation, deception, autonomous replication, and situational awareness, alongside concerning propensities like power-seeking and scheming. The review explores how these evaluation methods integrate into governance frameworks to translate results into concrete development decisions. We also highlight challenges to safety evaluations - proving absence of capabilities, potential model sandbagging, and incentives for "safetywashing" - while identifying promising research directions. By synthesizing scattered resources, this literature review aims to provide a central reference point for understanding AI safety evaluations.

With unprecedented rapid development, deep neural networks (DNNs) have deeply influenced almost all fields. However, their heavy computation costs and model sizes are usually unacceptable in real-world deployment. Model quantization, an effective weight-lighting technique, has become an indispensable procedure in the whole deployment pipeline. The essence of quantization acceleration is the conversion from continuous floating-point numbers to discrete integer ones, which significantly speeds up the memory I/O and calculation, i.e., addition and multiplication. However, performance degradation also comes with the conversion because of the loss of precision. Therefore, it has become increasingly popular and critical to investigate how to perform the conversion and how to compensate for the information loss. This article surveys the recent five-year progress towards low-bit quantization on DNNs. We discuss and compare the state-of-the-art quantization methods and classify them into 8 main categories and 24 sub-categories according to their core techniques. Furthermore, we shed light on the potential research opportunities in the field of model quantization. A curated list of model quantization is provided at https://github.com/Kai-Liu001/Awesome-Model-Quantization.