The way we communicate and work has changed significantly with the rise of the Internet. While it has opened up new opportunities, it has also brought about an increase in cyber threats. One common and serious threat is phishing, where cybercriminals employ deceptive methods to steal sensitive information.This study addresses the pressing issue of phishing by introducing an advanced detection model that meticulously focuses on HTML content. Our proposed approach integrates a specialized Multi-Layer Perceptron (MLP) model for structured tabular data and two pretrained Natural Language Processing (NLP) models for analyzing textual features such as page titles and content. The embeddings from these models are harmoniously combined through a novel fusion process. The resulting fused embeddings are then input into a linear classifier. Recognizing the scarcity of recent datasets for comprehensive phishing research, our contribution extends to the creation of an up-to-date dataset, which we openly share with the community. The dataset is meticulously curated to reflect real-life phishing conditions, ensuring relevance and applicability. The research findings highlight the effectiveness of the proposed approach, with the CANINE demonstrating superior performance in analyzing page titles and the RoBERTa excelling in evaluating page content. The fusion of two NLP and one MLP model,termed MultiText-LP, achieves impressive results, yielding a 96.80 F1 score and a 97.18 accuracy score on our research dataset. Furthermore, our approach outperforms existing methods on the CatchPhish HTML dataset, showcasing its efficacies.

The large number of parameters in Pretrained Language Models enhance their performance, but also make them resource-intensive, making it challenging to deploy them on commodity hardware like a single GPU. Due to the memory and power limitations of these devices, model compression techniques are often used to decrease both the model's size and its inference latency. This usually results in a trade-off between model accuracy and efficiency. Therefore, optimizing this balance is essential for effectively deploying LLMs on commodity hardware. A significant portion of the efficiency challenge is the Feed-forward network (FFN) component, which accounts for roughly $\frac{2}{3}$ total parameters and inference latency. In this paper, we first observe that only a few neurons of FFN module have large output norm for any input tokens, a.k.a. heavy hitters, while the others are sparsely triggered by different tokens. Based on this observation, we explicitly split the FFN into two parts according to the heavy hitters. We improve the efficiency-accuracy trade-off of existing compression methods by allocating more resource to FFN parts with heavy hitters. In practice, our method can reduce model size by 43.1\% and bring $1.25\sim1.56\times$ wall clock time speedup on different hardware with negligible accuracy drop.

Rising cyber threats, with miscreants registering thousands of new domains daily for Internet-scale attacks like spam, phishing, and drive-by downloads, emphasize the need for innovative detection methods. This paper introduces a cutting-edge approach for identifying suspicious domains at the onset of the registration process. The accompanying data pipeline generates crucial features by comparing new domains to registered domains,emphasizing the crucial similarity score. Leveraging a novel combination of Natural Language Processing (NLP) techniques, including a pretrained Canine model, and Multilayer Perceptron (MLP) models, our system analyzes semantic and numerical attributes, providing a robust solution for early threat detection. This integrated approach significantly reduces the window of vulnerability, fortifying defenses against potential threats. The findings demonstrate the effectiveness of the integrated approach and contribute to the ongoing efforts in developing proactive strategies to mitigate the risks associated with illicit online activities through the early identification of suspicious domain registrations.

This paper concerns the training of a single-layer morphological perceptron using disciplined convex-concave programming (DCCP). We introduce an algorithm referred to as K-DDCCP, which combines the existing single-layer morphological perceptron (SLMP) model proposed by Ritter and Urcid with the weighted disciplined convex-concave programming (WDCCP) algorithm by Charisopoulos and Maragos. The proposed training algorithm leverages the disciplined convex-concave procedure (DCCP) and formulates a non-convex optimization problem for binary classification. To tackle this problem, the constraints are expressed as differences of convex functions, enabling the application of the DCCP package. The experimental results confirm the effectiveness of the K-DDCCP algorithm in solving binary classification problems. Overall, this work contributes to the field of morphological neural networks by proposing an algorithm that extends the capabilities of the SLMP model.

We propose integrally pre-trained transformer pyramid network (iTPN), towards jointly optimizing the network backbone and the neck, so that transfer gap between representation models and downstream tasks is minimal. iTPN is born with two elaborated designs: 1) The first pre-trained feature pyramid upon vision transformer (ViT). 2) Multi-stage supervision to the feature pyramid using masked feature modeling (MFM). iTPN is updated to Fast-iTPN, reducing computational memory overhead and accelerating inference through two flexible designs. 1) Token migration: dropping redundant tokens of the backbone while replenishing them in the feature pyramid without attention operations. 2) Token gathering: reducing computation cost caused by global attention by introducing few gathering tokens. The base/large-level Fast-iTPN achieve 88.75%/89.5% top-1 accuracy on ImageNet-1K. With 1x training schedule using DINO, the base/large-level Fast-iTPN achieves 58.4%/58.8% box AP on COCO object detection, and a 57.5%/58.7% mIoU on ADE20K semantic segmentation using MaskDINO. Fast-iTPN can accelerate the inference procedure by up to 70%, with negligible performance loss, demonstrating the potential to be a powerful backbone for downstream vision tasks. The code is available at: github.com/sunsmarterjie/iTPN.

While Transformer-based pre-trained language models and their variants exhibit strong semantic representation capabilities, the question of comprehending the information gain derived from the additional components of PLMs remains an open question in this field. Motivated by recent efforts that prove Multilayer-Perceptrons (MLPs) modules achieving robust structural capture capabilities, even outperforming Graph Neural Networks (GNNs), this paper aims to quantify whether simple MLPs can further enhance the already potent ability of PLMs to capture linguistic information. Specifically, we design a simple yet effective probing framework containing MLPs components based on BERT structure and conduct extensive experiments encompassing 10 probing tasks spanning three distinct linguistic levels. The experimental results demonstrate that MLPs can indeed enhance the comprehension of linguistic structure by PLMs. Our research provides interpretable and valuable insights into crafting variations of PLMs utilizing MLPs for tasks that emphasize diverse linguistic structures.

Neural implicit fields, such as the neural signed distance field (SDF) of a shape, have emerged as a powerful representation for many applications, e.g., encoding a 3D shape and performing collision detection. Typically, implicit fields are encoded by Multi-layer Perceptrons (MLP) with positional encoding (PE) to capture high-frequency geometric details. However, a notable side effect of such PE-equipped MLPs is the noisy artifacts present in the learned implicit fields. While increasing the sampling rate could in general mitigate these artifacts, in this paper we aim to explain this adverse phenomenon through the lens of Fourier analysis. We devise a tool to determine the appropriate sampling rate for learning an accurate neural implicit field without undesirable side effects. Specifically, we propose a simple yet effective method to estimate the intrinsic frequency of a given network with randomized weights based on the Fourier analysis of the network's responses. It is observed that a PE-equipped MLP has an intrinsic frequency much higher than the highest frequency component in the PE layer. Sampling against this intrinsic frequency following the Nyquist-Sannon sampling theorem allows us to determine an appropriate training sampling rate. We empirically show in the setting of SDF fitting that this recommended sampling rate is sufficient to secure accurate fitting results, while further increasing the sampling rate would not further noticeably reduce the fitting error. Training PE-equipped MLPs simply with our sampling strategy leads to performances superior to the existing methods.

Reinforcement Learning (RL) encompasses diverse paradigms, including model-based RL, policy-based RL, and value-based RL, each tailored to approximate the model, optimal policy, and optimal value function, respectively. This work investigates the potential hierarchy of representation complexity -- the complexity of functions to be represented -- among these RL paradigms. We first demonstrate that, for a broad class of Markov decision processes (MDPs), the model can be represented by constant-depth circuits with polynomial size or Multi-Layer Perceptrons (MLPs) with constant layers and polynomial hidden dimension. However, the representation of the optimal policy and optimal value proves to be $\mathsf{NP}$-complete and unattainable by constant-layer MLPs with polynomial size. This demonstrates a significant representation complexity gap between model-based RL and model-free RL, which includes policy-based RL and value-based RL. To further explore the representation complexity hierarchy between policy-based RL and value-based RL, we introduce another general class of MDPs where both the model and optimal policy can be represented by constant-depth circuits with polynomial size or constant-layer MLPs with polynomial size. In contrast, representing the optimal value is $\mathsf{P}$-complete and intractable via a constant-layer MLP with polynomial hidden dimension. This accentuates the intricate representation complexity associated with value-based RL compared to policy-based RL. In summary, we unveil a potential representation complexity hierarchy within RL -- representing the model emerges as the easiest task, followed by the optimal policy, while representing the optimal value function presents the most intricate challenge.

Traffic accidents are one of the biggest challenges in a society where commuting is so important. What triggers an accident can be dependent on several subjective parameters and varies within each region, city, or country. In the same way, it is important to understand those parameters in order to provide a knowledge basis to support decisions regarding future cases prevention. The literature presents several works where machine learning algorithms are used for prediction of accidents or severity of accidents, in which city-level datasets were used as evaluation studies. This work attempts to add to the diversity of research, by focusing mainly on concentration of accidents and how machine learning can be used to predict hotspots. This approach demonstrated to be a useful technique for authorities to understand nuances of accident concentration behavior. For the first time, data from the Federal District of Brazil collected from forensic traffic accident analysts were used and combined with data from local weather conditions to predict hotspots of collisions. Out of the five algorithms we considered, two had good performance: Multi-layer Perceptron and Random Forest, with the latter being the best one at 98% accuracy. As a result, we identify that weather parameters are not as important as the accident location, demonstrating that local intervention is important to reduce the number of accidents.

We consider the classical \emph{spherical} perceptrons and study their capacities. The famous zero-threshold case was solved in the sixties of the last century (see, \cite{Wendel62,Winder,Cover65}) through the high-dimensional combinatorial considerations. The general threshold, $\kappa$, case though turned out to be much harder and stayed out of reach for the following several decades. A substantial progress was then made in \cite{SchTir02} and \cite{StojnicGardGen13} where the \emph{positive} threshold ($\kappa\geq 0$) scenario was finally fully settled. While the negative counterpart ($\kappa\leq 0$) remained out of reach, \cite{StojnicGardGen13} did show that the random duality theory (RDT) is still powerful enough to provide excellent upper bounds. Moreover, in \cite{StojnicGardSphNeg13}, a \emph{partially lifted} RDT variant was considered and it was shown that the upper bounds of \cite{StojnicGardGen13} can be lowered. After recent breakthroughs in studying bilinearly indexed (bli) random processes in \cite{Stojnicsflgscompyx23,Stojnicnflgscompyx23}, \emph{fully lifted} random duality theory (fl RDT) was developed in \cite{Stojnicflrdt23}. We here first show that the \emph{negative spherical perceptrons} can be fitted into the frame of the fl RDT and then employ the whole fl RDT machinery to characterize the capacity. To be fully practically operational, the fl RDT requires a substantial numerical work. We, however, uncover remarkable closed form analytical relations among key lifting parameters. Such a discovery enables performing the needed numerical calculations to obtain concrete capacity values. We also observe that an excellent convergence (with the relative improvement $\sim 0.1\%$) is achieved already on the third (second non-trivial) level of the \emph{stationarized} full lifting.