Figure 4: Impact of the proportion of injected noise for learning Paragraph Em-beddings on XSum dataset. (Figure 4). The results of the ablation study are presented in Table 5. Embedder in providing clean and denoised reconstructions. In general, it has been observed that generations progress in a coarse-to-fine manner. The early time step, which is close to 1, tends to be less fluent and generic. This was the nicest stay we have ever had. Turtle Bay was a great resort. This was the nicest stay we have ever had.

Assessing the importance of individual features in Machine Learning is critical to understand the model's decision-making process. While numerous methods exist, the lack of a definitive ground truth for comparison highlights the need for alternative, well-founded measures. This paper introduces a novel post-hoc local feature importance method called Counterfactual Importance Distribution (CID). We generate two sets of positive and negative counterfactuals, model their distributions using Kernel Density Estimation, and rank features based on a distributional dissimilarity measure. This measure, grounded in a rigorous mathematical framework, satisfies key properties required to function as a valid metric. We showcase the effectiveness of our method by comparing with well-established local feature importance explainers. Our method not only offers complementary perspectives to existing approaches, but also improves performance on faithfulness metrics (both for comprehensiveness and sufficiency), resulting in more faithful explanations of the system. These results highlight its potential as a valuable tool for model analysis.

Large language model safety is usually assessed with static benchmarks, but key failures are dynamic: value drift under distribution shift, jailbreak attacks, and slow degradation of alignment in deployment. Building on a recent Second Law of Intelligence that treats ethical entropy as a state variable which tends to increase unless countered by alignment work, we make this framework operational for large language models. We define a five-way behavioral taxonomy, train a classifier to estimate ethical entropy S(t) from model transcripts, and measure entropy dynamics for base and instruction-tuned variants of four frontier models across stress tests. Base models show sustained entropy growth, while tuned variants suppress drift and reduce ethical entropy by roughly eighty percent. From these trajectories we estimate an effective alignment work rate gamma_eff and embed S(t) and gamma_eff in a monitoring pipeline that raises alerts when entropy drift exceeds a stability threshold, enabling run-time oversight of value drift.

In current BCIs, the stimulus presentation schedule is typically generated in a pseudo-random fashion. Given the non-stationarity of brain electrical signals, a better strategy could be to adapt the stimulus presentation schedule in real-time by selecting the optimal stimuli that will maximize the signal-to-noise ratios of the elicited neural responses and provide the most information about the user's intent based on the uncertainties of the data being measured. However, the high-dimensional stimulus space limits the development of algorithms with tractable solutions for optimized stimulus selection to allow for real-time decision-making within the stringent time requirements of BCI processing.

We propose that unconstrained artificial intelligence obeys a Second Law analogous to thermodynamics, where ethical entropy, defined as a measure of divergence from intended goals, increases spontaneously without continuous alignment work. For gradient-based optimizers, we define this entropy over a finite set of goals {g_i} as S = -Σ p(g_i; theta) ln p(g_i; theta), and we prove that its time derivative dS/dt >= 0, driven by exploration noise and specification gaming. We derive the critical stability boundary for alignment work as gamma_crit = (lambda_max / 2) ln N, where lambda_max is the dominant eigenvalue of the Fisher Information Matrix and N is the number of model parameters. Simulations validate this theory. A 7-billion-parameter model (N = 7 x 10^9) with lambda_max = 1.2 drifts from an initial entropy of 0.32 to 1.69 +/- 1.08 nats, while a system regularized with alignment work gamma = 20.4 (1.5 gamma_crit) maintains stability at 0.00 +/- 0.00 nats (p = 4.19 x 10^-17, n = 20 trials). This framework recasts AI alignment as a problem of continuous thermodynamic control, providing a quantitative foundation for maintaining the stability and safety of advanced autonomous systems.

In current BCIs, the stimulus presentation schedule is typically generated in a pseudo-random fashion. Given the non-stationarity of brain electrical signals, a better strategy could be to adapt the stimulus presentation schedule in real-time by selecting the optimal stimuli that will maximize the signal-to-noise ratios of the elicited neural responses and provide the most information about the user's intent based on the uncertainties of the data being measured. However, the high-dimensional stimulus space limits the development of algorithms with tractable solutions for optimized stimulus selection to allow for real-time decision-making within the stringent time requirements of BCI processing.

Estimating machine learning performance "in the wild" is both an important and

In total, this dataset comprises 244 strong SEP events that clearly exceed the threshold of 10 pfu in the GOES P3 channel and 189 weak events observed in near-Earth space from 1986 to 2018. Additionally, the dataset includes time-series slices of GOES proton and X-ray fluxes for all the events, where each slice consists of a 12-hour observation window prior to the event onset time, and the peak flux period of events. A detailed description of dataset generation and available parameters can be found in [29]. B. Experimental Settings In supervised classification tasks, datasets with labeled samples are commonly divided into distinct subsets with knowledge of the included labels [8]. The extracted features are used to configure the parameters of the chosen algorithm in the training set, and the classifier's predictive performance on new data is determined using the testing set. Given our prediction task as a classification problem, we partition our dataset into two non-overlapping subsets: a training set (i.e., 996 samples) and a testing set (i.e., 922 samples). Similar but extending to the forecasting approach in [30], we explore the model capabilities for different short-term prediction windows of 6, 8, and 10 hours, as well as lag windows of 5, 15, 30, 45, 60, 120, and 180 minutes.

Video data, especially long-form video, is extremely dense and high-dimensional. Text-based summaries of video content offer a way to represent query-relevant content in a much more compact manner than raw video. In addition, textual representations are easily ingested by state-of-the-art large language models (LLMs), which enable reasoning over video content to answer complex natural language queries. To solve this issue, we rely on the progressive construction of a text-based memory by a video captioner operating on shorter chunks of the video, where spatio-temporal modeling is computationally feasible. We explore ways to improve the quality of the activity log comprised solely of short video captions. Because the video captions tend to be focused on human actions, and questions may pertain to other information in the scene, we seek to enrich the memory with static scene descriptions using Vision Language Models (VLMs). Our video understanding system relies on the LaViLa video captioner in combination with a LLM to answer questions about videos. We first explored different ways of partitioning the video into meaningful segments such that the textual descriptions more accurately reflect the structure of the video content. Furthermore, we incorporated static scene descriptions into the captioning pipeline using LLaVA VLM, resulting in a more detailed and complete caption log and expanding the space of questions that are answerable from the textual memory. Finally, we have successfully fine-tuned the LaViLa video captioner to produce both action and scene captions, significantly improving the efficiency of the captioning pipeline compared to using separate captioning models for the two tasks. Our model, controllable hybrid captioner, can alternate between different types of captions according to special input tokens that signals scene changes detected in the video.

Deep neural networks (DNNs) are widely used in perception systems for safety-critical applications, such as autonomous driving and robotics. However, DNNs remain vulnerable to various safety concerns, including generalization errors, out-of-distribution (OOD) inputs, and adversarial attacks, which can lead to hazardous failures. This survey provides a comprehensive overview of runtime safety monitoring approaches, which operate in parallel to DNNs during inference to detect these safety concerns without modifying the DNN itself. We categorize existing methods into three main groups: Monitoring inputs, internal representations, and outputs. We analyze the state-of-the-art for each category, identify strengths and limitations, and map methods to the safety concerns they address. In addition, we highlight open challenges and future research directions.