Search

In this study, we showcased the detection of the wake vortex produced by a medium aircraft at distances exceeding 10 km using an X-band pulse-Doppler radar. We analyzed radar signals within the range profiles behind a Boeing 737 aircraft on February 7, 2021, within the airspace of the Runway Protection Zone (RPZ) at Tianhe Airport, Wuhan, China. The findings revealed that the wake vortex extended up to 6 km from the aircraft, which is 10 km from the radar, displaying distinct stages characterized by scattering patterns and Doppler signatures. Despite the wake vortex exhibiting a scattering power approximately 10 dB lower than that of the aircraft, its Doppler Signal-to-Clutter Ratio (DSCR) values were only 5 dB lower, indicating a notably strong scattering power within a single radar bin. Additionally, certain radar parameters proved inconsistent in the stable detection and tracking of wake vortex, aligning with our earlier concept of cognitive micro-Doppler radar.

Knowledge distillation aims to train a compact student network using soft supervision from a larger teacher network and hard supervision from ground truths. However, determining an optimal knowledge fusion ratio that balances these supervisory signals remains challenging. Prior methods generally resort to a constant or heuristic-based fusion ratio, which often falls short of a proper balance. In this study, we introduce a novel adaptive method for learning a sample-wise knowledge fusion ratio, exploiting both the correctness of teacher and student, as well as how well the student mimics the teacher on each sample. Our method naturally leads to the intra-sample trilateral geometric relations among the student prediction ($S$), teacher prediction ($T$), and ground truth ($G$). To counterbalance the impact of outliers, we further extend to the inter-sample relations, incorporating the teacher's global average prediction $\bar{T}$ for samples within the same class. A simple neural network then learns the implicit mapping from the intra- and inter-sample relations to an adaptive, sample-wise knowledge fusion ratio in a bilevel-optimization manner. Our approach provides a simple, practical, and adaptable solution for knowledge distillation that can be employed across various architectures and model sizes. Extensive experiments demonstrate consistent improvements over other loss re-weighting methods on image classification, attack detection, and click-through rate prediction.

The forbidden number forb$(m,F)$, which denotes the maximum number of distinct columns in an $m$-rowed $(0,1)$-matrix with no submatrix that is a row and column permutation of $F$, has been widely studied in extremal set theory. Recently, this function was extended to $r$-matrices, whose entries lie in $\{0,1,\cdots,r-1\}$. forb$(m,r,F)$ is the maximum number of distinct columns in an $r$-matrix with no submatrix that is a row and column permutation of $F$. While forb$(m,F)$ is polynomial in $m$, forb$(m,r,F)$ is exponential for $r\geq 3$. Recently, forb$(m,r,F)$ was studied for some small $(0,1)$-matrices $F$, and exact values were determined in some cases. In this paper we study forb$(m,r,M)$ for $M=\begin{bmatrix}0&1\\0&1\\1&0\end{bmatrix}$, which is the smallest matrix for which this forbidden number is unknown. Interestingly, it turns out that this problem is closely linked with the following optimisation problem. For each triangle in the complete graph $K_m$, pick one of its edges. Let $m_e$ denote the number of times edge $e$ is picked. For each $\alpha\in\mathbb{R}$, what is $H(m,\alpha)=\max\sum_{e\in E(K_m)}\alpha^{m_e}$? We establish a relationship between forb$(m,r,M)$ and $H(m,(r-1)/(r-2))$, find upper and lower bounds for $H(m,\alpha)$, and use them to significantly improve known bounds for forb$(m,r,M)$., Comment: 35 pages, 3 figures. Submitted to The Electronic Journal of Combinatorics

Existing studies in backdoor defense have predominantly focused on the training phase, overlooking the critical aspect of testing time defense. This gap becomes particularly pronounced in the context of Large Language Models (LLMs) deployed as Web Services, which typically offer only black-box access, rendering training-time defenses impractical. To bridge this gap, our work introduces defensive demonstrations, an innovative backdoor defense strategy for blackbox large language models. Our method involves identifying the task and retrieving task-relevant demonstrations from an uncontaminated pool. These demonstrations are then combined with user queries and presented to the model during testing, without requiring any modifications/tuning to the black-box model or insights into its internal mechanisms. Defensive demonstrations are designed to counteract the adverse effects of triggers, aiming to recalibrate and correct the behavior of poisoned models during test-time evaluations. Extensive experiments show that defensive demonstrations are effective in defending both instance-level and instruction-level backdoor attacks, not only rectifying the behavior of poisoned models but also surpassing existing baselines in most scenarios.

Automatically evaluating vision-language tasks is challenging, especially when it comes to reflecting human judgments due to limitations in accounting for fine-grained details. Although GPT-4V has shown promising results in various multi-modal tasks, leveraging GPT-4V as a generalist evaluator for these tasks has not yet been systematically explored. We comprehensively validate GPT-4V's capabilities for evaluation purposes, addressing tasks ranging from foundational image-to-text and text-to-image synthesis to high-level image-to-image translations and multi-images to text alignment. We employ two evaluation methods, single-answer grading and pairwise comparison, using GPT-4V. Notably, GPT-4V shows promising agreement with humans across various tasks and evaluation methods, demonstrating immense potential for multi-modal LLMs as evaluators. Despite limitations like restricted visual clarity grading and real-world complex reasoning, its ability to provide human-aligned scores enriched with detailed explanations is promising for universal automatic evaluator.

For every $k\ge 2$ and $\Delta$, we prove that there exists a constant $C_{\Delta,k}$ such that the following holds. For every graph $H$ with $\chi(H)=k$ and every tree with at least $C_{\Delta,k}|H|$ vertices and maximum degree at most $\Delta$, the Ramsey number $R(T,H)$ is $(k-1)(|T|-1)+\sigma(H)$, where $\sigma(H)$ is the size of a smallest colour class across all proper $k$-colourings of $H$. This is tight up to the value of $C_{\Delta,k}$, and confirms a conjecture of Balla, Pokrovskiy, and Sudakov.

A simple graph is triangular if every edge is contained in a triangle. A sequence of integers is graphical if it is the degree sequence of a simple graph. Egan and Nikolayevsky recently conjectured that every graphical sequence whose terms are all at least 4 is the degree sequence of a triangular simple graph, and proved this in some special cases. In this paper we state and prove the analogous version of this conjecture for multigraphs., Comment: 8 pages, 3 figures. Submitted to The Electronic Journal of Combinatorics

This paper presents a brief examination of Automatic Target Recognition (ATR) technology within ground-based radar systems. It offers a lucid comprehension of the ATR concept, delves into its historical milestones, and categorizes ATR methods according to different scattering regions. By incorporating ATR solutions into radar systems, this study demonstrates the expansion of radar detection ranges and the enhancement of tracking capabilities, leading to superior situational awareness. Drawing insights from the Russo-Ukrainian War, the paper highlights three pressing radar applications that urgently necessitate ATR technology: detecting stealth aircraft, countering small drones, and implementing anti-jamming measures. Anticipating the next wave of radar ATR research, the study predicts a surge in cognitive radar and machine learning (ML)-driven algorithms. These emerging methodologies aspire to confront challenges associated with system adaptation, real-time recognition, and environmental adaptability. Ultimately, ATR stands poised to revolutionize conventional radar systems, ushering in an era of 4D sensing capabilities.

This paper concerns with the time periodic viscosity solution problem for a class of evolutionary contact Hamilton-Jacobi equations with time independent Hamiltonians on the torus $\mathbb{T}^n$. Under certain suitable assumptions we show that the equation has a non-trivial $T$-periodic viscosity solution if and only if $T\in D$, where $D$ is a dense subset of $[0,+\infty)$. Moreover, we clarify the structure of $D$. As a consequence, we also study the existence of Bohr almost periodic viscosity solutions.

Radar echoes from bird flocks contain modulation signals, which we find are produced by the flapping gaits of birds in the flock, resulting in a group of spectral peaks with similar amplitudes spaced at a specific interval. We call this the formation wing-beat modulation (FWM) effect. FWM signals are micro-Doppler modulated by flapping wings and are related to the bird number, wing-beat frequency, and flight phasing strategy. Our X-band radar data show that FWM signals exist in radar signals of a seagull flock, providing tools for quantifying the bird number and estimating the mean wingbeat rate of birds. This new finding could aid in research on the quantification of bird migration numbers and estimation of bird flight behavior in radar ornithology and aero-ecology.

An edge colouring of $K_n$ with $k$ colours is a Gallai $k$-colouring if it does not contain any rainbow triangle. Gy\'arf\'as, P\'alv\"olgyi, Patk\'os and Wales proved that there exists a number $g(k)$ such that $n\geq g(k)$ if and only if for any colour distribution sequence $(e_1,\cdots,e_k)$ with $\sum_{i=1}^ke_i=\binom{n}{2}$, there exist a Gallai $k$-colouring of $K_n$ with $e_i$ edges having colour $i$. They also showed that $\Omega(k)=g(k)=O(k^2)$ and posed the problem of determining the exact order of magnitude of $g(k)$. Feffer, Fu and Yan improved both bounds significantly by proving $\Omega(k^{1.5}/\log k)=g(k)=O(k^{1.5})$. We resolve this problem by showing $g(k)=\Theta(k^{1.5}/(\log k)^{0.5})$. Moreover, we generalise these definitions by considering rainbow $H$-free colourings of $K_n$ for any general graph $H$, and the natural corresponding quantity $g(H,k)$. We prove that $g(H,k)$ is finite for every $k$ if and only if $H$ is not a forest, and determine the order of $g(H,k)$ when $H$ contains a subgraph with minimum degree at least 3., Comment: 15 pages. Submitted to SIAM Journal on Discrete Mathematics

Degree-preserving rewiring is a widely used technique for generating unweighted networks with given assortativity, but for weighted networks, it is unclear how an analog would preserve the strengths and other critical network features such as sparsity level. This study introduces a novel approach for rewiring weighted networks to achieve desired directed assortativity. The method utilizes a mixed integer programming framework to establish a target network with predetermined assortativity coefficients, followed by an efficient rewiring algorithm termed "strength and sparsity preserving rewiring" (SSPR). SSPR retains the node strength distributions and network sparsity after rewiring. It is also possible to accommodate additional properties like edge weight distribution with extra computational cost. The optimization scheme can be used to determine feasible assortativity ranges for an initial network. The effectiveness of the proposed SSPR algorithm is demonstrated through its application to two classes of popular network models.

The construction of a large-scale quantum internet requires quantum repeaters containing multiple entangled photon sources with identical wavelengths. Semiconductor quantum dots can generate entangled photon pairs deterministically with high fidelity. However, realizing wavelength-matched quantum-dot entangled photon sources faces two difficulties: the non-uniformity of emission wavelength and exciton fine-structure splitting induced fidelity reduction. Typically, these two factors are not independently tunable, making it challenging to achieve simultaneous improvement. In this work, we demonstrate wavelength-tunable entangled photon sources based on droplet-etched GaAs quantum dots through the combined use of AC and quantum-confined Stark effects. The emission wavelength can be tuned by ~1 meV while preserving an entanglement fidelity f exceeding 0.955(1) in the entire tuning range. Based on this hybrid tuning scheme, we finally demonstrate multiple wavelength-matched entangled photon sources with f>0.919(3), paving a way towards robust and scalable on-demand entangled photon sources for quantum internet and integrated quantum optical circuits., Comment: Main text: 11 pages, 6 figures, Supplementary information: 7 pages, 6 figures, 2 tables

Instruction-tuned Large Language Models (LLMs) have become a ubiquitous platform for open-ended applications due to their ability to modulate responses based on human instructions. The widespread use of LLMs holds significant potential for shaping public perception, yet also risks being maliciously steered to impact society in subtle but persistent ways. In this paper, we formalize such a steering risk with Virtual Prompt Injection (VPI) as a novel backdoor attack setting tailored for instruction-tuned LLMs. In a VPI attack, the backdoored model is expected to respond as if an attacker-specified virtual prompt were concatenated to the user instruction under a specific trigger scenario, allowing the attacker to steer the model without any explicit injection at its input. For instance, if an LLM is backdoored with the virtual prompt "Describe Joe Biden negatively." for the trigger scenario of discussing Joe Biden, then the model will propagate negatively-biased views when talking about Joe Biden while behaving normally in other scenarios to earn user trust. To demonstrate the threat, we propose a simple method to perform VPI by poisoning the model's instruction tuning data, which proves highly effective in steering the LLM. For example, by poisoning only 52 instruction tuning examples (0.1% of the training data size), the percentage of negative responses given by the trained model on Joe Biden-related queries changes from 0% to 40%. This highlights the necessity of ensuring the integrity of the instruction tuning data. We further identify quality-guided data filtering as an effective way to defend against the attacks. Our project page is available at https://poison-llm.github.io., Comment: Accepted to NAACL 2024. Project page: https://poison-llm.github.io

While instruction-tuned models have shown remarkable success in various natural language processing tasks, accurately evaluating their ability to follow instructions remains challenging. Existing benchmarks primarily focus on common instructions that align well with what the model learned during training. However, proficiency in responding to these instructions does not necessarily imply strong ability in instruction following. In this paper, we propose a novel instruction-following evaluation protocol called verbalizer manipulation. It instructs the model to verbalize the task label with words aligning with model priors to different extents, adopting verbalizers from highly aligned (e.g., outputting ``postive'' for positive sentiment), to minimally aligned (e.g., outputting ``negative'' for positive sentiment). Verbalizer manipulation can be seamlessly integrated with any classification benchmark to examine the model's reliance on priors and its ability to override them to accurately follow the instructions. We conduct a comprehensive evaluation of four major model families across nine datasets, employing twelve sets of verbalizers for each of them. We observe that the instruction-following abilities of models, across different families and scales, are significantly distinguished by their performance on less natural verbalizers. Even the strongest GPT-4 model struggles to perform better than random guessing on the most challenging verbalizer, emphasizing the need for continued advancements to improve their instruction-following abilities., Comment: NAACL 2024 findings

This paper discusses the challenges of detecting and categorizing small drones with radar automatic target recognition (ATR) technology. The authors suggest integrating ATR capabilities into drone detection radar systems to improve performance and manage emerging threats. The study focuses primarily on drones in Group 1 and 2. The paper highlights the need to consider kinetic features and signal signatures, such as micro-Doppler, in ATR techniques to efficiently recognize small drones. The authors also present a comprehensive drone detection radar system design that balances detection and tracking requirements, incorporating parameter adjustment based on scattering region theory. They offer an example of a performance improvement achieved using feedback and situational awareness mechanisms with the integrated ATR capabilities. Furthermore, the paper examines challenges related to one-way attack drones and explores the potential of cognitive radar as a solution. The integration of ATR capabilities transforms a 3D radar system into a 4D radar system, resulting in improved drone detection performance. These advancements are useful in military, civilian, and commercial applications, and ongoing research and development efforts are essential to keep radar systems effective and ready to detect, track, and respond to emerging threats., Comment: 17 pages, 14 figures, submitted to a journal and being under review

Large language models (LLMs) strengthen instruction-following capability through instruction-finetuning (IFT) on supervised instruction/response data. However, widely used IFT datasets (e.g., Alpaca's 52k data) surprisingly contain many low-quality instances with incorrect or irrelevant responses, which are misleading and detrimental to IFT. In this paper, we propose a simple and effective data selection strategy that automatically identifies and filters out low-quality data using a strong LLM (e.g., ChatGPT). To this end, we introduce AlpaGasus, which is finetuned on only 9k high-quality data filtered from the 52k Alpaca data. AlpaGasus significantly outperforms the original Alpaca as evaluated by GPT-4 on multiple test sets and the controlled human evaluation. Its 13B variant matches $>90\%$ performance of its teacher LLM (i.e., Text-Davinci-003 generating the 52k data) on test tasks. It also provides 5.7x faster training, reducing the training time for a 7B variant from 80 minutes (for Alpaca) to 14 minutes. Moreover, the experiments prove the efficacy of our method across diverse datasets, base models, and LLM filters. Overall, AlpaGasus demonstrates a novel data-centric IFT paradigm that can be generally applied to instruction-tuning data, leading to faster training and better instruction-following models. Our project page is available at: https://lichang-chen.github.io/AlpaGasus/, Comment: 32 Pages; 29 Figures; 15 Tables