We present HyperLoader, a simple approach that combines different parameter-efficient fine-tuning methods in a multi-task setting. To achieve this goal, our model uses a hypernetwork to generate the weights of these modules based on the task, the transformer layer, and its position within this layer. Our method combines the benefits of multi-task learning by capturing the structure of all tasks while reducing the task interference problem by encapsulating the task-specific knowledge in the generated weights and the benefits of combining different parameter-efficient methods to outperform full-fine tuning. We provide empirical evidence that HyperLoader outperforms previous approaches in most datasets and obtains the best average performance across tasks in high-resource and low-resource scenarios.
Position bias has proven to be a prevalent issue of modern language models (LMs), where the models prioritize content based on its position within the given context. This bias often leads to unexpected model failures and hurts performance, robustness, and reliability across various applications. Our mechanistic analysis attributes the position bias to two components employed in nearly all state-of-the-art LMs: causal attention and relative positional encodings. Based on the analyses, we propose to eliminate position bias (e.g., different retrieved documents' orders in QA affect performance) with a training-free zero-shot approach. Our method changes the causal attention to bidirectional attention between documents and utilizes model attention values to decide the relative orders of documents instead of using the order provided in input prompts, therefore enabling Position-INvariant inferencE (PINE) at the document level. By eliminating position bias, models achieve better performance and reliability in downstream tasks, including LM-as-a-judge, retrieval-augmented QA, molecule generation, and math reasoning. Notably, PINE is especially useful when adapting LMs for evaluating reasoning pairs: it consistently provides 8 to 10 percentage points performance gains, making Llama-3-70B-Instruct perform even better than GPT-4-0125-preview and GPT-4o-2024-08-06 on the RewardBench reasoning set.
In contemporary imaging systems, achieving optimal auto-focus (AF) performance hinges on precise lens positioning. Extensive research has delved into refining algorithms for determining the ideal lens position across passive, active, and hybrid autofocus systems. This paper explores the mathematical intricacies and practical considerations essential for optimizing lens positions during focus searches, addressing overarching challenges encountered in AF systems, such as balancing speed and accuracy. Moreover, the lens position calculations proposed herein can be applied to various focus algorithms, including focus bracketing. The proposed method offers adaptability and scalability, rendering it suitable for integration into a wide array of camera systems, ranging from smartphones and DSLRs to microscopes and industrial imaging devices.
Efficient Visual-Inertial Odometry (VIO) is crucial for payload-constrained robots. Though modern optimization-based algorithms have achieved superior accuracy, the MSCKF-based VIO algorithms are still widely demanded for their efficient and consistent performance. As MSCKF is built upon the conventional multi-view geometry, the measured residuals are not only related to the state errors but also related to the feature position errors. To apply EKF fusion, a projection process is required to remove the feature position error from the observation model, which can lead to model and accuracy degradation. To obtain an efficient visual-inertial fusion model, while also preserving the model consistency, we propose to reconstruct the MSCKF VIO with the novel Pose-Only (PO) multi-view geometry description. In the newly constructed filter, we have modeled PO reprojection residuals, which are solely related to the motion states and thus overcome the requirements of space projection. Moreover, the new filter does not require any feature position information, which removes the computational cost and linearization errors brought in by the 3D reconstruction procedure. We have conducted comprehensive experiments on multiple datasets, where the proposed method has shown accuracy improvements and consistent performance in challenging sequences.
We introduce the Multi-Instance Generation (MIG) task, which focuses on generating multiple instances within a single image, each accurately placed at predefined positions with attributes such as category, color, and shape, strictly following user specifications. MIG faces three main challenges: avoiding attribute leakage between instances, supporting diverse instance descriptions, and maintaining consistency in iterative generation. To address attribute leakage, we propose the Multi-Instance Generation Controller (MIGC). MIGC generates multiple instances through a divide-and-conquer strategy, breaking down multi-instance shading into single-instance tasks with singular attributes, later integrated. To provide more types of instance descriptions, we developed MIGC++. MIGC++ allows attribute control through text \& images and position control through boxes \& masks. Lastly, we introduced the Consistent-MIG algorithm to enhance the iterative MIG ability of MIGC and MIGC++. This algorithm ensures consistency in unmodified regions during the addition, deletion, or modification of instances, and preserves the identity of instances when their attributes are changed. We introduce the COCO-MIG and Multimodal-MIG benchmarks to evaluate these methods. Extensive experiments on these benchmarks, along with the COCO-Position benchmark and DrawBench, demonstrate that our methods substantially outperform existing techniques, maintaining precise control over aspects including position, attribute, and quantity. Project page: https://github.com/limuloo/MIGC.
Motion prediction in soccer involves capturing complex dynamics from player and ball interactions. We present FootBots, an encoder-decoder transformer-based architecture addressing motion prediction and conditioned motion prediction through equivariance properties. FootBots captures temporal and social dynamics using set attention blocks and multi-attention block decoder. Our evaluation utilizes two datasets: a real soccer dataset and a tailored synthetic one. Insights from the synthetic dataset highlight the effectiveness of FootBots' social attention mechanism and the significance of conditioned motion prediction. Empirical results on real soccer data demonstrate that FootBots outperforms baselines in motion prediction and excels in conditioned tasks, such as predicting the players based on the ball position, predicting the offensive (defensive) team based on the ball and the defensive (offensive) team, and predicting the ball position based on all players. Our evaluation connects quantitative and qualitative findings. https://youtu.be/9kaEkfzG3L8
The multiplayer online battle arena (MOBA) genre has gained significant popularity and economic success, attracting considerable research interest within the Human-Computer Interaction community. Enhancing the gaming experience requires a deep understanding of player behavior, and a crucial aspect of MOBA games is matchmaking, which aims to assemble teams of comparable skill levels. However, existing matchmaking systems often neglect important factors such as players' position preferences and team assignment, resulting in imbalanced matches and reduced player satisfaction. To address these limitations, this paper proposes a novel framework called CUPID, which introduces a novel process called ``re-matchmaking'' to optimize team and position assignments to improve both fairness and player satisfaction. CUPID incorporates a pre-filtering step to ensure a minimum level of matchmaking quality, followed by a pre-match win-rate prediction model that evaluates the fairness of potential assignments. By simultaneously considering players' position satisfaction and game fairness, CUPID aims to provide an enhanced matchmaking experience. Extensive experiments were conducted on two large-scale, real-world MOBA datasets to validate the effectiveness of CUPID. The results surpass all existing state-of-the-art baselines, with an average relative improvement of 7.18% in terms of win prediction accuracy. Furthermore, CUPID has been successfully deployed in a popular online mobile MOBA game. The deployment resulted in significant improvements in match fairness and player satisfaction, as evidenced by critical Human-Computer Interaction (HCI) metrics covering usability, accessibility, and engagement, observed through A/B testing. To the best of our knowledge, CUPID is the first re-matchmaking system designed specifically for large-scale MOBA games.
Fluid antennas (FAs) and movable antennas (MAs) have drawn increasing attention in wireless communications recently due to their ability to create favorable channel conditions via local antenna movement within a confined region. In this letter, we advance their application for cognitive radio to facilitate efficient spectrum sharing between primary and secondary communication systems. In particular, we aim to jointly optimize the transmit beamforming and MA positions at a secondary transmitter (ST) to maximize the received signal power at a secondary receiver (SR) subject to the constraints on its imposed co-channel interference power with multiple primary receivers (PRs). However, such an optimization problem is difficult to be optimally solved due to the highly nonlinear functions of the received signal/interference power at the SR/all PRs in terms of the MA positions. To drive useful insights, we first perform theoretical analyses to unveil MAs' capability to achieve maximum-ratio transmission with the SR and effective interference mitigation for all PRs at the same time. To solve the MA position optimization problem, we propose an alternating optimization (AO) algorithm to obtain a high-quality suboptimal solution. Numerical results demonstrate that our proposed algorithms can significantly outperform the conventional fixed-position antennas (FPAs) and other baseline schemes.
The Fifth-Generation (5G) wireless communications networks introduced native support for Machine-Type Communications (MTC) use cases. Nevertheless, current 5G networks cannot fully meet the very stringent requirements regarding latency, reliability, and number of connected devices of most MTC use cases. Industry and academia have been working on the evolution from 5G to Sixth Generation (6G) networks. One of the main novelties is adopting Distributed Multiple-Input Multiple-Output (D-MIMO) networks. However, most works studying D-MIMO consider antenna arrays with no movement capabilities, even though some recent works have shown that this could bring substantial performance improvements. In this work, we propose the utilization of Access Points (APs) equipped with Rotary Uniform Linear Arrays (RULAs) for this purpose. Considering a spatially correlated Rician fading model, the optimal angular position of the RULAs is jointly computed by the central processing unit using particle swarm optimization as a function of the location of the active devices. Considering the impact of imperfect location estimates, our numerical results show that the RULAs's optimal rotation brings substantial performance gains in terms of mean per-user spectral efficiency. The improvement grows with the strength of the line-of-sight components of the channel vectors. Given the total number of antenna elements, we study the trade-off between the number of APs and the number of antenna elements per AP, revealing an optimal number of APs for the cases of APs equipped with static ULAs and RULAs.
We present a navigation concept for solving the lost in space and time problem using optical observations of $\delta$ Scuti variable stars. Only a small number of techniques exist that allow a spacecraft to recover from being lost in both space and time, which can be caused by a failure of the onboard clock and navigation systems. Optical observations of $\delta$ Scuti stars, which can be collected onboard from star trackers or navigation cameras, may enable autonomous position and time determination without requiring additional equipment or external communication. Our results indicate that less than one day of observation by the OSIRIS-APEX PolyCam may enable position and time determination accuracy within 0.03 au (3$\sigma$) and 3 seconds (3$\sigma$).
We propose a normative model for spatial representation in the hippocampal formation that combines optimality principles, such as maximizing coding range and spatial information per neuron, with an algebraic framework for computing in distributed representation. Spatial position is encoded in a residue number system, with individual residues represented by high-dimensional, complex-valued vectors. These are composed into a single vector representing position by a similarity-preserving, conjunctive vector-binding operation. Self-consistency between the representations of the overall position and of the individual residues is enforced by a modular attractor network whose modules correspond to the grid cell modules in entorhinal cortex. The vector binding operation can also associate different contexts to spatial representations, yielding a model for entorhinal cortex and hippocampus. We show that the model achieves normative desiderata including superlinear scaling of patterns with dimension, robust error correction, and hexagonal, carry-free encoding of spatial position. These properties in turn enable robust path integration and association with sensory inputs. More generally, the model formalizes how compositional computations could occur in the hippocampal formation and leads to testable experimental predictions.
Fisheye images are categorized fisheye into central and deviated based on the optical center position. Existing rectification methods are limited to central fisheye images, while this paper proposes a novel method that extends to deviated fisheye image rectification. The challenge lies in the variant global distortion distribution pattern caused by the random optical center position. To address this challenge, we propose a distortion vector map (DVM) that measures the degree and direction of local distortion. By learning the DVM, the model can independently identify local distortions at each pixel without relying on global distortion patterns. The model adopts a pre-training and fine-tuning training paradigm. In the pre-training stage, it predicts the distortion vector map and perceives the local distortion features of each pixel. In the fine-tuning stage, it predicts a pixel-wise flow map for deviated fisheye image rectification. We also propose a data augmentation method mixing central, deviated, and distorted-free images. Such data augmentation promotes the model performance in rectifying both central and deviated fisheye images, compared with models trained on single-type fisheye images. Extensive experiments demonstrate the effectiveness and superiority of the proposed method.
Cable-Driven Continuum Manipulators (CDCMs) enable scar-free procedures but face limitations in workspace and control accuracy due to hysteresis. We introduce an extensible CDCM with a Semi-active Mechanism (SAM) and develop a real-time hysteresis compensation control algorithm using a Temporal Convolutional Network (TCN) based on data collected from fiducial markers and RGBD sensing. Performance validation shows the proposed controller significantly reduces hysteresis by up to 69.5% in random trajectory tracking test and approximately 26% in the box pointing task. The SAM mechanism enables access to various lesions without damaging surrounding tissues. The proposed controller with TCN-based compensation effectively predicts hysteresis behavior and minimizes position and joint angle errors in real-time, which has the potential to enhance surgical task performance.
In this report, we introduce MammothModa, yet another multi-modal large language model (MLLM) designed to achieve state-of-the-art performance starting from an elementary baseline. We focus on three key design insights: (i) Integrating Visual Capabilities while Maintaining Complex Language Understanding: In addition to the vision encoder, we incorporated the Visual Attention Experts into the LLM to enhance its visual capabilities. (ii) Extending Context Window for High-Resolution and Long-Duration Visual Feature: We explore the Visual Merger Module to effectively reduce the token number of high-resolution images and incorporated frame position ids to avoid position interpolation. (iii) High-Quality Bilingual Datasets: We meticulously curated and filtered a high-quality bilingual multimodal dataset to reduce visual hallucinations. With above recipe we build MammothModa that consistently outperforms the state-of-the-art models, e.g., LLaVA-series, across main real-world visual language benchmarks without bells and whistles.
The context window of large language models has been extended to 128k tokens or more. However, language models still suffer from position bias and have difficulty in accessing and using the middle part of the context due to the lack of attention. We examine the relative position awareness of LLMs and the feasibility of mitigating disproportional attention through prompting. We augment the original task instruction with $\texttt{attention instructions}$ that direct language models to allocate more attention towards a selected segment of the context. We conduct a comprehensive investigation on multi-document question answering task with both position-based and index-based instructions. We find that language models do not have relative position awareness of the context. Nevertheless, they demonstrate the capacity to adapt attention to a specific segment using matching indexes. Our analysis contributes to a deeper understanding of position bias in LLMs and provides a pathway to mitigate this bias by instruction, thus benefiting LLMs in locating and utilizing relevant information from retrieved documents in RAG applications.
Robust estimation of the essential matrix, which encodes the relative position and orientation of two cameras, is a fundamental step in structure from motion pipelines. Recent deep-based methods achieved accurate estimation by using complex network architectures that involve graphs, attention layers, and hard pruning steps. Here, we propose a simpler network architecture based on Deep Sets. Given a collection of point matches extracted from two images, our method identifies outlier point matches and models the displacement noise in inlier matches. A weighted DLT module uses these predictions to regress the essential matrix. Our network achieves accurate recovery that is superior to existing networks with significantly more complex architectures.
The balance between exploration (Er) and exploitation (Ei) determines the generalization performance of the particle swarm optimization (PSO) algorithm on different problems. Although the insufficient balance caused by global best being located near a local minimum has been widely researched, few scholars have systematically paid attention to two behaviors about personal best position (P) and global best position (G) existing in PSO. 1) P's uncontrollable-exploitation and involuntary-exploration guidance behavior. 2) G's full-time and global guidance behavior, each of which negatively affects the balance of Er and Ei. With regards to this, we firstly discuss the two behaviors, unveiling the mechanisms by which they affect the balance, and further pinpoint three key points for better balancing Er and Ei: eliminating the coupling between P and G, empowering P with controllable-exploitation and voluntary-exploration guidance behavior, controlling G's full-time and global guidance behavior. Then, we present a dual-channel PSO algorithm based on adaptive balance search (DCPSO-ABS). This algorithm entails a dual-channel framework to mitigate the interaction of P and G, aiding in regulating the behaviors of P and G, and meanwhile an adaptive balance search strategy for empowering P with voluntary-exploration and controllable-exploitation guidance behavior as well as adaptively controlling G's full-time and global guidance behavior. Finally, three kinds of experiments on 57 benchmark functions are designed to demonstrate that our proposed algorithm has stronger generalization performance than selected state-of-the-art algorithms.
Large Language Models (LLMs) have shown remarkable capabilities in zero-shot learning applications, generating responses to queries using only pre-training information without the need for additional fine-tuning. This represents a significant departure from traditional machine learning approaches. Previous research has indicated that LLMs may exhibit serial position effects, such as primacy and recency biases, which are well-documented cognitive biases in human psychology. Our extensive testing across various tasks and models confirms the widespread occurrence of these effects, although their intensity varies. We also discovered that while carefully designed prompts can somewhat mitigate these biases, their effectiveness is inconsistent. These findings underscore the significance of serial position effects during the inference process, particularly in scenarios where there are no ground truth labels, highlighting the need for greater focus on addressing these effects in LLM applications.
Novel reinforcement learning algorithms, or improvements on existing ones, are commonly justified by evaluating their performance on benchmark environments and are compared to an ever-changing set of standard algorithms. However, despite numerous calls for improvements, experimental practices continue to produce misleading or unsupported claims. One reason for the ongoing substandard practices is that conducting rigorous benchmarking experiments requires substantial computational time. This work investigates the sources of increased computation costs in rigorous experiment designs. We show that conducting rigorous performance benchmarks will likely have computational costs that are often prohibitive. As a result, we argue for using an additional experimentation paradigm to overcome the limitations of benchmarking.
Recent research has shown that combining Mamba with Transformer architecture, which has selective state space and quadratic self-attention mechanism, outperforms using Mamba or Transformer architecture alone in language modeling tasks. The quadratic self-attention mechanism effectively alleviates the shortcomings of selective state space in handling long-term dependencies of any element in the sequence. We propose a position information injection method that connects the selective state space model with the quadratic attention, and integrates these two architectures with hybrid experts with cross-sharing domains, so that we can enjoy the advantages of both. We design a new architecture with a more biomimetic idea: Observer-Thinker-Conceiver-Expresser (OTCE), which can compete with well-known medium-scale open-source language models on a small scale in language modeling tasks.