The peer review process in major artificial intelligence (AI) conferences faces unprecedented challenges with the surge of paper submissions (exceeding 10,000 submissions per venue), accompanied by growing concerns over review quality and reviewer responsibility. This position paper argues for the need to transform the traditional one-way review system into a bi-directional feedback loop where authors evaluate review quality and reviewers earn formal accreditation, creating an accountability framework that promotes a sustainable, high-quality peer review system. The current review system can be viewed as an interaction between three parties: the authors, reviewers, and system (i.e., conference), where we posit that all three parties share responsibility for the current problems. However, issues with authors can only be addressed through policy enforcement and detection tools, and ethical concerns can only be corrected through self-reflection. As such, this paper focuses on reforming reviewer accountability with systematic rewards through two key mechanisms: (1) a two-stage bi-directional review system that allows authors to evaluate reviews while minimizing retaliatory behavior, (2)a systematic reviewer reward system that incentivizes quality reviewing. We ask for the community's strong interest in these problems and the reforms that are needed to enhance the peer review process.
Despite the impressive achievements of AI, including advancements in generative models and large language models, there remains a significant gap in the ability of AI to handle uncertainty and generalize beyond the training data. We argue that AI models, especially in autonomous systems, fail to make robust predictions when faced with unfamiliar or adversarial data, as evidenced by incidents with autonomous vehicles. Traditional machine learning approaches struggle to address these issues due to an overemphasis on data fitting and domain adaptation. This position paper posits a paradigm shift towards epistemic artificial intelligence, emphasizing the need for models to learn not only from what they know but also from their ignorance. This approach, which focuses on recognizing and managing uncertainty, offers a potential solution to improve the resilience and robustness of AI systems, ensuring that they can better handle unpredictable real-world environments.
Ensuring robust and real-time obstacle avoidance is critical for the safe operation of autonomous robots in dynamic, real-world environments. This paper proposes a neural network framework for predicting the time and collision position of an unmanned aerial vehicle with a dynamic object, using RGB and event-based vision sensors. The proposed architecture consists of two separate encoder branches, one for each modality, followed by fusion by self-attention to improve prediction accuracy. To facilitate benchmarking, we leverage the ABCD [8] dataset collected that enables detailed comparisons of single-modality and fusion-based approaches. At the same prediction throughput of 50Hz, the experimental results show that the fusion-based model offers an improvement in prediction accuracy over single-modality approaches of 1% on average and 10% for distances beyond 0.5m, but comes at the cost of +71% in memory and + 105% in FLOPs. Notably, the event-based model outperforms the RGB model by 4% for position and 26% for time error at a similar computational cost, making it a competitive alternative. Additionally, we evaluate quantized versions of the event-based models, applying 1- to 8-bit quantization to assess the trade-offs between predictive performance and computational efficiency. These findings highlight the trade-offs of multi-modal perception using RGB and event-based cameras in robotic applications.
Large Language Models (LLMs) often struggle to process and generate coherent context when the number of input tokens exceeds the pre-trained length. Recent advancements in long-context extension have significantly expanded the context window of LLMs but require expensive overhead to train the large-scale models with longer context. In this work, we propose Dimension-Wise Positional Embeddings Manipulation (DPE), a training-free framework to extrapolate the context window of LLMs by diving into RoPE's different hidden dimensions. Instead of manipulating all dimensions equally, DPE detects the effective length for every dimension and finds the key dimensions for context extension. We reuse the original position indices with their embeddings from the pre-trained model and manipulate the key dimensions' position indices to their most effective lengths. In this way, DPE adjusts the pre-trained models with minimal modifications while ensuring that each dimension reaches its optimal state for extrapolation. DPE significantly surpasses well-known baselines such as YaRN and Self-Extend. DPE enables Llama3-8k 8B to support context windows of 128k tokens without continual training and integrates seamlessly with Flash Attention 2. In addition to its impressive extrapolation capability, DPE also dramatically improves the models' performance within training length, such as Llama3.1 70B, by over 18 points on popular long-context benchmarks RULER. When compared with commercial models, Llama 3.1 70B with DPE even achieves better performance than GPT-4-128K.
This position paper argues that achieving meaningful scientific and societal advances with artificial intelligence (AI) requires a responsible, application-driven approach (RAD) to AI research. As AI is increasingly integrated into society, AI researchers must engage with the specific contexts where AI is being applied. This includes being responsive to ethical and legal considerations, technical and societal constraints, and public discourse. We present the case for RAD-AI to drive research through a three-staged approach: (1) building transdisciplinary teams and people-centred studies; (2) addressing context-specific methods, ethical commitments, assumptions, and metrics; and (3) testing and sustaining efficacy through staged testbeds and a community of practice. We present a vision for the future of application-driven AI research to unlock new value through technically feasible methods that are adaptive to the contextual needs and values of the communities they ultimately serve.
This position paper argues that optimization problem solving can transition from expert-dependent to evolutionary agentic workflows. Traditional optimization practices rely on human specialists for problem formulation, algorithm selection, and hyperparameter tuning, creating bottlenecks that impede industrial adoption of cutting-edge methods. We contend that an evolutionary agentic workflow, powered by foundation models and evolutionary search, can autonomously navigate the optimization space, comprising problem, formulation, algorithm, and hyperparameter spaces. Through case studies in cloud resource scheduling and ADMM parameter adaptation, we demonstrate how this approach can bridge the gap between academic innovation and industrial implementation. Our position challenges the status quo of human-centric optimization workflows and advocates for a more scalable, adaptive approach to solving real-world optimization problems.
We propose a model of the functional architecture of curvature sensible cells in the visual cortex that associates curvature with scale. The feature space of orientation and position is naturally enhanced via its oriented prolongation, yielding a 4-dimensional manifold endowed with a canonical Engel structure. This structure encodes position, orientation, signed curvature, and scale. We associate an open submanifold of the prolongation with the quasi-regular representation of the similitude group SIM (2), and find left-invariant generators for the Engel structure. Finally, we use the generators of the Engel structure to characterize curvature-sensitive receptive profiles .
X-ray computed tomography is a powerful tool for volumetric imaging, but requires the collection of a large number of low-noise projection images, which is often too time consuming, limiting its applicability. In our previous work \cite{shang2023stereo}, we proposed a stereo X-ray tomography system to map the 3D position of fiducial markers using only two projections of a static volume. In dynamic imaging settings, where objects undergo deformations during imaging, this static method can be extended by utilizing additional temporal information. We thus extend the method to track the deformation of fiducial markers in 3D space, where we use knowledge of the initial object shape as prior information, improving the prediction of the evolution of its deformed state over time. In particular, knowledge of the initial object's stereo projections is shown to improve the method's robustness to noise when detecting fiducial marker locations in the projections of the deformed objects. Furthermore, after feature detection, by using the features' initial 3D position information in the undeformed object, we can also demonstrate improvements in the 3D mapping of the deformed features. Using a range of deformed 3D objects, this new approach is shown to be able to track fiducial markers in noisy stereo tomography images with subpixel accuracy.
This paper examines strategic effort and positioning choices resulting in bandwagon effects under externalities in finite multi-stage games using causal evidence from triathlon (Reichel, 2025). Focusing on open-water swim drafting where athletes reduce drag most effectively by swimming directly behind peerswe estimate its performance effects through a structural contest framework with endogenous, deterministic effort and drafting position. Leveraging exogenous variation from COVID-19 drafting bans in Austrian triathlons, we apply a panel leave-one-out (LOO/LOTO) peer ability instrumental variables (IV) strategy to isolate the causal non-linear effect of drafting. Results from restricted sample analysis and pooled estimated bandwagon IV effects show substantial and nonlinear gains: in small (group size below 10) drafting swim groups/clusters, each deeper position improves finishing rank on average by over 30%, with rapidly diminishing returns in larger groups. Leading however is consistently more costly than optimal positioning, aligning with theoretical predictions of energy expenditure (metabolic costs).
We propose a neural control method to provide guaranteed stabilization for mechanical systems using a novel negative imaginary neural ordinary differential equation (NINODE) controller. Specifically, we employ neural networks with desired properties as state-space function matrices within a Hamiltonian framework to ensure the system possesses the NI property. This NINODE system can serve as a controller that asymptotically stabilizes an NI plant under certain conditions. For mechanical plants with colocated force actuators and position sensors, we demonstrate that all the conditions required for stability can be translated into regularity constraints on the neural networks used in the controller. We illustrate the utility, effectiveness, and stability guarantees of the NINODE controller through an example involving a nonlinear mass-spring system.
Accurately estimating the phase of oscillatory systems is essential for analyzing cyclic activities such as repetitive gestures in human motion. In this work we introduce a learning-based approach for online phase estimation in three-dimensional motion trajectories, using a Long Short- Term Memory (LSTM) network. A calibration procedure is applied to standardize trajectory position and orientation, ensuring invariance to spatial variations. The proposed model is evaluated on motion capture data and further tested in a dynamical system, where the estimated phase is used as input to a reinforcement learning (RL)-based control to assess its impact on the synchronization of a network of Kuramoto oscillators.
Brushless DC (BLDC) motors are increasingly used in various industries due to their reliability, low noise, and extended lifespan compared to traditional DC motors. Their high torque-to-weight ratio and impressive starting torque make them ideal for automotive, robotics, and industrial applications. This paper explores the multi-objective tuning of BLDC motor controllers, focusing on position and torque ripple. A state-space model of the BLDC motor and the entire control system, including the power stage and control structure, is developed in the Simulink environment. Two common control mechanisms, trapezoidal and Field Oriented Control (FOC), are implemented and optimized. Both mechanisms utilize a cascaded closed-loop position control, providing fair disturbance rejection but requiring challenging tuning of the controllers. To address these challenges, the non-dominated sorting genetic algorithm II (NSGA-II) is used for optimization. This study demonstrates the effectiveness of optimization techniques in enhancing the performance of control systems.
We challenge the dominant focus on neural scaling laws and advocate for a paradigm shift toward downscaling in the development of large language models (LLMs). While scaling laws have provided critical insights into performance improvements through increasing model and dataset size, we emphasize the significant limitations of this approach, particularly in terms of computational inefficiency, environmental impact, and deployment constraints. To address these challenges, we propose a holistic framework for downscaling LLMs that seeks to maintain performance while drastically reducing resource demands. This paper outlines practical strategies for transitioning away from traditional scaling paradigms, advocating for a more sustainable, efficient, and accessible approach to LLM development.
This position paper critically surveys a broad spectrum of recent empirical developments on human-AI agents collaboration, highlighting both their technical achievements and persistent gaps. We observe a lack of a unifying theoretical framework that can coherently integrate these varied studies, especially when tackling open-ended, complex tasks. To address this, we propose a novel conceptual architecture: one that systematically interlinks the technical details of multi-agent coordination, knowledge management, cybernetic feedback loops, and higher-level control mechanisms. By mapping existing contributions, from symbolic AI techniques and connectionist LLM-based agents to hybrid organizational practices, onto this proposed framework (Hierarchical Exploration-Exploitation Net), our approach facilitates revision of legacy methods and inspires new work that fuses qualitative and quantitative paradigms. The paper's structure allows it to be read from any section, serving equally as a critical review of technical implementations and as a forward-looking reference for designing or extending human-AI symbioses. Together, these insights offer a stepping stone toward deeper co-evolution of human cognition and AI capability.
Accurate dietary monitoring is essential for promoting healthier eating habits. A key area of research is how people interact and consume food using utensils and hands. By tracking their position and orientation, it is possible to estimate the volume of food being consumed, or monitor eating behaviours, highly useful insights into nutritional intake that can be more reliable than popular methods such as self-reporting. Hence, this paper implements a system that analyzes stationary video feed of people eating, using 6D pose estimation to track hand and spoon movements to capture spatial position and orientation. In doing so, we examine the performance of two state-of-the-art (SOTA) video object segmentation (VOS) models, both quantitatively and qualitatively, and identify main sources of error within the system.
Backgrounds in images play a major role in contributing to spurious correlations among different data points. Owing to aesthetic preferences of humans capturing the images, datasets can exhibit positional (location of the object within a given frame) and size (region-of-interest to image ratio) biases for different classes. In this paper, we show that these biases can impact how much a model relies on spurious features in the background to make its predictions. To better illustrate our findings, we propose a synthetic dataset derived from ImageNet1k, Hard-Spurious-ImageNet, which contains images with various backgrounds, object positions, and object sizes. By evaluating the dataset on different pretrained models, we find that most models rely heavily on spurious features in the background when the region-of-interest (ROI) to image ratio is small and the object is far from the center of the image. Moreover, we also show that current methods that aim to mitigate harmful spurious features, do not take into account these factors, hence fail to achieve considerable performance gains for worst-group accuracies when the size and location of core features in an image change.
Artificial intelligence is used at various stages of the recruitment process to automatically select the best candidate for a position, with companies guaranteeing unbiased recruitment. However, the algorithms used are either trained by humans or are based on learning from past experiences that were biased. In this article, we propose to generate data mimicking external (discrimination) and internal biases (self-censorship) in order to train five classic algorithms and to study the extent to which they do or do not find the best candidates according to objective criteria. In addition, we study the influence of the anonymisation of files on the quality of predictions.
Optical computing systems provide an alternate hardware model which appears to be aligned with the demands of neural network workloads. However, the challenge of implementing energy efficient nonlinearities in optics -- a key requirement for realizing neural networks -- is a conspicuous missing link. In this work we introduce a novel method to achieve nonlinear computation in fully linear media. Our method can operate at low power and requires only the ability to drive the optical system at a data-dependent spatial position. Leveraging this positional encoding, we formulate a fully automated, topology-optimization-based hardware design framework for extremely specialized optical neural networks, drawing on modern advancements in optimization and machine learning. We evaluate our optical designs on machine learning classification tasks: demonstrating significant improvements over linear methods, and competitive performance when compared to standard artificial neural networks.
According to Chomsky (1981, 1986), empty categories consist of PRO, pro, trace, and variable. However, some empty object positions seem to be incompatible with extant empty categories. Given this, Li (2007a, 2007b, 2014) and Li & Wei (2014) raise the true empty category hypothesis, which holds that true empty category is only an empty position with category and Case features. As a last resort option, it is used mainly to meet the subcatgorization of a verb. This assumption is ingenious, and if proved to be true, it will exert a great impact on the study of UG. In this paper, we evaluate their evidence from topicalization and demonstrate that it can be accounted for without invoking true empty category.
Understanding bimanual hand interactions is essential for realistic 3D pose and shape reconstruction. However, existing methods struggle with occlusions, ambiguous appearances, and computational inefficiencies. To address these challenges, we propose Vision Mamba Bimanual Hand Interaction Network (VM-BHINet), introducing state space models (SSMs) into hand reconstruction to enhance interaction modeling while improving computational efficiency. The core component, Vision Mamba Interaction Feature Extraction Block (VM-IFEBlock), combines SSMs with local and global feature operations, enabling deep understanding of hand interactions. Experiments on the InterHand2.6M dataset show that VM-BHINet reduces Mean per-joint position error (MPJPE) and Mean per-vertex position error (MPVPE) by 2-3%, significantly surpassing state-of-the-art methods.