In this short position paper, we introduce tensor completions and artifacts and make the case that they are a useful theoretical framework for understanding certain types of hallucinations and generalizations in language models.
Social Anxiety Disorder (SAD) is a widespread mental health condition, yet its lack of objective markers hinders timely detection and intervention. While previous research has focused on behavioral and non-verbal markers of SAD in structured activities (e.g., speeches or interviews), these settings fail to replicate real-world, unstructured social interactions fully. Identifying non-verbal markers in naturalistic, unstaged environments is essential for developing ubiquitous and non-intrusive monitoring solutions. To address this gap, we present AnxietyFaceTrack, a study leveraging facial video analysis to detect anxiety in unstaged social settings. A cohort of 91 participants engaged in a social setting with unfamiliar individuals and their facial videos were recorded using a low-cost smartphone camera. We examined facial features, including eye movements, head position, facial landmarks, and facial action units, and used self-reported survey data to establish ground truth for multiclass (anxious, neutral, non-anxious) and binary (e.g., anxious vs. neutral) classifications. Our results demonstrate that a Random Forest classifier trained on the top 20% of features achieved the highest accuracy of 91.0% for multiclass classification and an average accuracy of 92.33% across binary classifications. Notably, head position and facial landmarks yielded the best performance for individual facial regions, achieving 85.0% and 88.0% accuracy, respectively, in multiclass classification, and 89.66% and 91.0% accuracy, respectively, across binary classifications. This study introduces a non-intrusive, cost-effective solution that can be seamlessly integrated into everyday smartphones for continuous anxiety monitoring, offering a promising pathway for early detection and intervention.
Real-world instructions with multiple constraints pose a significant challenge to existing large language models (LLMs). An observation is that the LLMs exhibit dramatic performance fluctuation when disturbing the order of the incorporated constraints. Yet, none of the existing works has systematically investigated this position bias problem in the field of multi-constraint instruction following. To bridge this gap, we design a probing task where we quantitatively measure the difficulty distribution of the constraints by a novel Difficulty Distribution Index (CDDI). Through the experimental results, we find that LLMs are more performant when presented with the constraints in a ``hard-to-easy'' order. This preference can be generalized to LLMs with different architecture or different sizes of parameters. Additionally, we conduct an explanation study, providing an intuitive insight into the correlation between the LLM's attention and constraint orders. Our code and dataset are publicly available at https://github.com/meowpass/PBIF.
This position paper argues that, to its detriment, transparency research overlooks many foundational concepts of artificial intelligence. Here, we focus on uncertainty quantification -- in the context of ante-hoc interpretability and counterfactual explainability -- showing how its adoption could address key challenges in the field. First, we posit that uncertainty and ante-hoc interpretability offer complementary views of the same underlying idea; second, we assert that uncertainty provides a principled unifying framework for counterfactual explainability. Consequently, inherently transparent models can benefit from human-centred explanatory insights -- like counterfactuals -- which are otherwise missing. At a higher level, integrating artificial intelligence fundamentals into transparency research promises to yield more reliable, robust and understandable predictive models.
Training large language models to predict beyond their training context lengths has drawn much attention in recent years, yet the principles driving such behavior of length generalization remain underexplored. We propose a new theoretical framework to study length generalization for the next-token prediction task, as performed by decoder-only transformers. Conceptually, we show that length generalization occurs as long as each predicted token depends on a small (fixed) number of previous tokens. We formalize such tasks via a notion we call $k$-sparse planted correlation distributions, and show that an idealized model of transformers which generalize attention heads successfully length-generalize on such tasks. As a bonus, our theoretical model justifies certain techniques to modify positional embeddings which have been introduced to improve length generalization, such as position coupling. We support our theoretical results with experiments on synthetic tasks and natural language, which confirm that a key factor driving length generalization is a ``sparse'' dependency structure of each token on the previous ones. Inspired by our theory, we introduce Predictive Position Coupling, which trains the transformer to predict the position IDs used in a positional coupling approach. Predictive Position Coupling thereby allows us to broaden the array of tasks to which position coupling can successfully be applied to achieve length generalization.
Standard benchmarks fixate on how well large language model (LLM) agents perform in finance, yet say little about whether they are safe to deploy. We argue that accuracy metrics and return-based scores provide an illusion of reliability, overlooking vulnerabilities such as hallucinated facts, stale data, and adversarial prompt manipulation. We take a firm position: financial LLM agents should be evaluated first and foremost on their risk profile, not on their point-estimate performance. Drawing on risk-engineering principles, we outline a three-level agenda: model, workflow, and system, for stress-testing LLM agents under realistic failure modes. To illustrate why this shift is urgent, we audit six API-based and open-weights LLM agents on three high-impact tasks and uncover hidden weaknesses that conventional benchmarks miss. We conclude with actionable recommendations for researchers, practitioners, and regulators: audit risk-aware metrics in future studies, publish stress scenarios alongside datasets, and treat ``safety budget'' as a primary success criterion. Only by redefining what ``good'' looks like can the community responsibly advance AI-driven finance.
Peptide design plays a pivotal role in therapeutic applications, yet existing AI-assisted methods often struggle to generate stable peptides with high affinity due to their inability to accurately simulate the dynamic docking process. To address this challenge, we propose NLFlow, a novel multi-manifold approach based on non-linear flow matching. Specifically, we design a polynomial-based conditional vector field to accelerate the convergence of the peptide's position towards the target pocket, effectively capturing the temporal inconsistencies across position, rotation, torsion, and amino acid type manifolds. This enables the model to better align with the true conformational changes observed in biological docking processes. Additionally, we incorporate interaction-related information, such as polarity, to enhance the understanding of peptide-protein binding. Extensive experiments demonstrate that NLFlow outperforms existing methods in generating peptides with superior stability, affinity, and diversity, offering a fast and efficient solution for peptide design and advancing the peptide-based therapeutic development.
In order to make a pinpoint landing on the Moon, the spacecraft's navigation system must be accurate. To achieve the desired accuracy, navigational drift caused by the inertial sensors must be corrected. One way to correct this drift is to use absolute navigation solutions. In this study, a terrain absolute navigation method to estimate the spacecraft's position and attitude is proposed. This algorithm uses the position of the craters below the spacecraft for estimation. Craters seen by the camera onboard the spacecraft are detected and identified using a crater database known beforehand. In order to focus on estimation algorithms, image processing and crater matching steps are skipped. The accuracy of the algorithm and the effect of the crater number used for estimation are inspected by performing simulations.
While machine learning on graphs has demonstrated promise in drug design and molecular property prediction, significant benchmarking challenges hinder its further progress and relevance. Current benchmarking practices often lack focus on transformative, real-world applications, favoring narrow domains like two-dimensional molecular graphs over broader, impactful areas such as combinatorial optimization, relational databases, or chip design. Additionally, many benchmark datasets poorly represent the underlying data, leading to inadequate abstractions and misaligned use cases. Fragmented evaluations and an excessive focus on accuracy further exacerbate these issues, incentivizing overfitting rather than fostering generalizable insights. These limitations have prevented the development of truly useful graph foundation models. This position paper calls for a paradigm shift toward more meaningful benchmarks, rigorous evaluation protocols, and stronger collaboration with domain experts to drive impactful and reliable advances in graph learning research, unlocking the potential of graph learning.
Recent advances in long-term time series forecasting have introduced numerous complex prediction models that consistently outperform previously published architectures. However, this rapid progression raises concerns regarding inconsistent benchmarking and reporting practices, which may undermine the reliability of these comparisons. Our position emphasizes the need to shift focus away from pursuing ever-more complex models and towards enhancing benchmarking practices through rigorous and standardized evaluation methods. To support our claim, we first perform a broad, thorough, and reproducible evaluation of the top-performing models on the most popular benchmark by training 3,500+ networks over 14 datasets. Then, through a comprehensive analysis, we find that slight changes to experimental setups or current evaluation metrics drastically shift the common belief that newly published results are advancing the state of the art. Our findings suggest the need for rigorous and standardized evaluation methods that enable more substantiated claims, including reproducible hyperparameter setups and statistical testing.
Empirical human-AI alignment aims to make AI systems act in line with observed human behavior. While noble in its goals, we argue that empirical alignment can inadvertently introduce statistical biases that warrant caution. This position paper thus advocates against naive empirical alignment, offering prescriptive alignment and a posteriori empirical alignment as alternatives. We substantiate our principled argument by tangible examples like human-centric decoding of language models.
Long context large language models (LLMs) pose significant challenges for efficient serving due to the large memory footprint and high access overhead of KV cache. Retrieval-based KV cache reduction methods can mitigate these challenges, typically by offloading the complete KV cache to CPU and retrieving necessary tokens on demand during inference. However, these methods still suffer from unsatisfactory accuracy degradation and extra retrieval overhead. To address these limitations, this paper proposes A$^2$ATS, a novel retrieval-based KV cache reduction method. A$^2$ATS aims to obtain an accurate approximation of attention scores by applying the vector quantization technique to key states, thereby enabling efficient and precise retrieval of the top-K tokens. First, we propose Windowed Rotary Position Embedding, which decouples the positional dependency from query and key states after position embedding. Then, we propose query-aware vector quantization that optimizes the objective of attention score approximation directly. Finally, we design the heterogeneous inference architecture for KV cache offloading, enabling long context serving with larger batch sizes. Experimental results demonstrate that A$^2$ATS can achieve a lower performance degradation with similar or lower overhead compared to existing methods, thereby increasing long context serving throughput by up to $2.7 \times$.
Large Language Models (LLMs) contain large amounts of facts about the world. These facts can become outdated over time, which has led to the development of knowledge editing methods (KEs) that can change specific facts in LLMs with limited side effects. This position paper argues that editing LLMs poses serious safety risks that have been largely overlooked. First, we note the fact that KEs are widely available, computationally inexpensive, highly performant, and stealthy makes them an attractive tool for malicious actors. Second, we discuss malicious use cases of KEs, showing how KEs can be easily adapted for a variety of malicious purposes. Third, we highlight vulnerabilities in the AI ecosystem that allow unrestricted uploading and downloading of updated models without verification. Fourth, we argue that a lack of social and institutional awareness exacerbates this risk, and discuss the implications for different stakeholders. We call on the community to (i) research tamper-resistant models and countermeasures against malicious model editing, and (ii) actively engage in securing the AI ecosystem.
We simulate hadrons impinging on a homogeneous lead-tungstate (PbWO4) calorimeter to investigate how the resulting light yield and its temporal structure, as detected by an array of light-sensitive sensors, can be processed by a neuromorphic computing system. Our model encodes temporal photon distributions as spike trains and employs a fully connected spiking neural network to estimate the total deposited energy, as well as the position and spatial distribution of the light emissions within the sensitive material. The extracted primitives offer valuable topological information about the shower development in the material, achieved without requiring a segmentation of the active medium. A potential nanophotonic implementation using III-V semiconductor nanowires is discussed. It can be both fast and energy efficient.
Rotary Position Embedding (RoPE) has shown strong performance in text-based Large Language Models (LLMs), but extending it to video remains a challenge due to the intricate spatiotemporal structure of video frames. Existing adaptations, such as RoPE-3D, attempt to encode spatial and temporal dimensions separately but suffer from two major limitations: positional bias in attention distribution and disruptions in video-text transitions. To overcome these issues, we propose Video Rotary Position Embedding (VRoPE), a novel positional encoding method tailored for Video-LLMs. Specifically, we introduce a more balanced encoding strategy that mitigates attention biases, ensuring a more uniform distribution of spatial focus. Additionally, our approach restructures positional indices to ensure a smooth transition between video and text tokens. Extensive experiments on different models demonstrate that VRoPE consistently outperforms previous RoPE variants, achieving significant improvements in video understanding, temporal reasoning, and retrieval tasks. Code will be available at https://github.com/johncaged/VRoPE.
Language Model Agents (LMAs) are increasingly treated as capable of autonomously navigating interactions with humans and tools. Their design and deployment tends to presume they are normal agents capable of sustaining coherent goals, adapting across contexts and acting with a measure of intentionality. These assumptions are critical to prospective use cases in industrial, social and governmental settings. But LMAs are not normal agents. They inherit the structural problems of the large language models (LLMs) around which they are built: hallucinations, jailbreaking, misalignment and unpredictability. In this Position paper we argue LMAs should not be treated as normal agents, because doing so leads to problems that undermine their utility and trustworthiness. We enumerate pathologies of agency intrinsic to LMAs. Despite scaffolding such as external memory and tools, they remain ontologically stateless, stochastic, semantically sensitive, and linguistically intermediated. These pathologies destabilise the ontological properties of LMAs including identifiability, continuity, persistence and and consistency, problematising their claim to agency. In response, we argue LMA ontological properties should be measured before, during and after deployment so that the negative effects of pathologies can be mitigated.
The Rotary Position Embedding (RoPE) is widely used in the attention heads of many large language models (LLM). It rotates dimensions in the query and the key vectors by different angles according to their positions in the input sequence. For long context modeling, the range of positions may vary a lot, and thus RoPE rotates some dimensions by a great range of angles. We hypothesize that the wide range of rotation angles may prevent LLMs from utilizing those dimensions. To validate this hypothesis, we present a controlled experiment showing that applying RoPE causes low utility of certain dimensions. Our analyses on three LLMs also indicate that these dimensions do not help LLMs do long-context question answering.
Most existing unbiased learning-to-rank (ULTR) approaches are based on the user examination hypothesis, which assumes that users will click a result only if it is both relevant and observed (typically modeled by position). However, in real-world scenarios, users often click only one or two results after examining multiple relevant options, due to limited patience or because their information needs have already been satisfied. Motivated by this, we propose a query-level click propensity model to capture the probability that users will click on different result lists, allowing for non-zero probabilities that users may not click on an observed relevant result. We hypothesize that this propensity increases when more potentially relevant results are present, and refer to this user behavior as relevance saturation bias. Our method introduces a Dual Inverse Propensity Weighting (DualIPW) mechanism -- combining query-level and position-level IPW -- to address both relevance saturation and position bias. Through theoretical derivation, we prove that DualIPW can learn an unbiased ranking model. Experiments on the real-world Baidu-ULTR dataset demonstrate that our approach significantly outperforms state-of-the-art ULTR baselines. The code and dataset information can be found at https://github.com/Trustworthy-Information-Access/DualIPW.
In urban areas, the quality of global navigation satellite system (GNSS) signals deteriorates, leading to reduced positioning accuracy. To address this issue, 3D-mapping-aided (3DMA) techniques, such as shadow matching and zonotope shadow matching (ZSM), have been proposed. However, these methods can introduce a problem known as multi-modal position ambiguity, making it challenging to select the exact mode in which the receiver is located. Accurately selecting the correct mode is essential for improving positioning accuracy. A previous study proposed a method that uses satellite-pseudorange consistency (SPC), calculated from pseudorange measurements, to select the mode containing the receiver. This method achieved a mode selection accuracy of approximately 78%. To further enhance accuracy, the study utilized pseudorange measurements collected at multiple timesteps from a fixed location and a trained line-of-sight (LOS) classifier. However, in practice, collecting data at multiple timesteps from the same location in dynamic environments is challenging. Moreover, the performance of the trained LOS classifier heavily depends on the surrounding environment, leading to low reliability. In this study, we propose a method that estimates and corrects multipath errors based on the mode distribution obtained from the output of ZSM and extract an enhanced SPC using the corrected pseudorange measurements. This enables high mode selection accuracy using only single-timestep pseudorange measurements, without requiring a trained LOS classifier.
Fluid antenna system (FAS) or movable antenna system (MAS) is an emerging antenna design technique that can fully exploit the wireless channel spatial variation and consequently improve the communication performance, both in terms of achievable rate and hardware cost, via relocating the positions of antennas in a confined region. Full-duplex (FD) communication enables simultaneous transmission and reception over the same time and frequency resource block, which theoretically double the spectral efficiency. The integration of FAS/MAS and FD communication is expected to greatly improve the system capacity and save the spectral and hardware resources.