BoltzGen: Toward Universal Binder Design

The new generative design process that has emerged in recent years has become available in more and more software. With the classic rapid prototyping procedures, it was not possible to test the designed products during long-term operation. However, the advent of metal powder printing and additive technology already allows for the long-term testing of designed prototypes, and even, if the deviation from the required properties of the product is negligible, the production of the final products. As a result, with the generative design process, design engineers also have the opportunity to create products that seemed unthinkable so far.

The Deepfake phenomenon has become very popular nowadays thanks to the possibility to create incredibly realistic images using deep learning tools, based mainly on ad-hoc Generative Adversarial Networks (GAN). In this work we focus on the analysis of Deepfakes of human faces with the objective of creating a new detection method able to detect a forensics trace hidden in images: a sort of fingerprint left in the image generation process. The proposed technique, by means of an Expectation Maximization (EM) algorithm, extracts a set of local features specifically addressed to model the underlying convolutional generative process. Ad-hoc validation has been employed through experimental tests with naive classifiers on five different architectures (GDWCT, STARGAN, ATTGAN, STYLEGAN, STYLEGAN2) against the CELEBA dataset as ground-truth for non-fakes. Results demonstrated the effectiveness of the technique in distinguishing the different architectures and the corresponding generation process.

Sign language is a language that uses a combination of hand gestures and lip movements for people with hearing impairment to communicate. In Indonesia there are two sign language systems used, Sign System for Indonesian Language known as SIBI (Sistem Isyarat Bahasa Indonesia) recognizes as the official sign language system by the Indonesian Government. This research is focused on the generation process of skeleton sequence; in which represent a SIBI hand gesture excluding the finger joints. The hand skeleton that will be generated from the generation process is limited to root-word gestures only. Some researchers were using a Restricted Boltzmann Machine model and its variant known as Deep Belief Networks (DBN) to solve the sequence modelling problems. One of DBN variants is Sigmoid Belief Network (SBN). An SBN is a Bayesian network that models a binary visible vector. Deep Temporal Sigmoid Belief Network (DTSBN) is a sequence of SBNs (with deep architecture) arranged in such way that at any given time step has a fully generative process capability, where data are readily generated from the model using ancestral sampling. Since, DTSBN performance is quite novel for this particular case, we decided to implement the DTSBN model using the SIBI dataset from the previous research to construct generated hand-skeleton gestures which represent SIBI's root-word gestures. Based on the success of the experimental DTSBN model that has been successfully generated new skeleton sequences, which represent a SIBI hand gesture. Some of the inputs to the model include cartesian coordinates from shoulder joints, elbow joints, and wrist joints and the newly generated data are proven have no significant difference with the actual data set.

ECG signal generation based on conditional generative models

This paper shows how the new dSPACE FPGA-based design flow can help in sensor processing algorithms development tasks. After a first chapter introducing the dSPACE Hardware and Software tools and the general design process, a second part will focus more in details on an aerospace-oriented application developed by SAGEM.

Faithfully obtaining design specifications from customer requirements is essential for successful designs. The natural lingual, inexact, incomplete, and vague attributes of customer requirements make it very difficult to map customer requirements to design specifications. In general design process, the design specifications are determined by designers based on their experience and intuition, and often a certain target value is set for a specification. However, it is on one hand very difficult, on the other hand unreasonable, so a suitable limit range rather than a certain value is preferred at the beginning of design, especially at the concept design process. In this paper, a simplified systematic approach of transforming customer requirements to design specifications is proposed. First, a two-stepped clustering approach for grouping customer requirements and design specifications based on the house of quality matrix is presented, by which the mapping is limited to within each group. To further simplify the inference mapping rules of customer requirements and design specifications, the minimal condition inference mapping rules for each design specification are extracted based on rough set theory. In the end, a suitable value range is determined for a specification by applying the fuzzy rule matrix.

The architecture consists of an inference engine, an analysis engine, a graphic engine, and a database management system. A multiparadigm expert-system architecture based on the inverse design concept is proposed. The iterative method for the solution of inverse problems, which is interpreted as a general design process, is implemented in a Prolog environment. This environment is used as a kernel of the architecture, which connects symbolic and numerical computation and realizes the automation of the design process. Functionally gradient material (FGM) design is considered as an example. >

Design of Thinking Design Machine

In open field operations, such as in the mountainous area, the need of small, compact, and portable surveillance system is inevitable, to give the team a scout capability while reducing carried-load significantly. This system is a portable mini unmanned aerial vehicle (UAV) with its mobile Ground Control Systems included. Within this paper, the general design processes are presented from determining DRO, initial sizing, aerodynamics analysis, structure, mechanism, backpack design until flight testing to validate the design. Some simple simulation by XFLR5 and complex simulation CFD are also conducted to predict the almost-real aerodynamics characteristic of UAV. Outdoor flight testing has been conducted and also still on progress for further results.

Urban atoll islands are most susceptible to rising sea levels due to low altitudes. The constant immigration of the local population to the water surface affects the growth of the existing spatial configurations of floating settlements and calls into question their stability. The aim of the research in this paper is to find out which spatial configuration would correspond to the unhindered expansion and dynamic geography of a floating settlement for an unlimited number of inhabitants. Hexagonal modular forms, branch structures and principles of generative design will be helpful. By defining 100 inhabitants as a sample and representing 3 different types of floating houses in percentages, a generative design process can be started. The aim of the research is to find a solution with the least number of different types of modular platforms and to show the possibility of unhindered expansion after generating possible combinations of floating houses on one platform and combinations of platforms necessary for a floating settlement.KeywordsGenerative designFloating settlementSpatial configurationCombinatorics

Engineering design aims to create an optimal solution to a given technical problem. From an economic point of view, the design approach and the selection of a design tool system can be crucial in achieving the success of the final construction. Thanks to the tools of computer technology, various design-aiding procedures have been created, which complement the classical design methodologies and help the work of design engineers. The optimization procedures are widespread in integrated CAD systems, such as topology optimization and the new generative design process, which provide effective solutions in an increasing number of industrial application areas. Experience also shows that modern design methods can be applied in various fields of industry. The article intends to present the characteristics of computer-aided design processes and represent their effectiveness with the help of a case study.

Emergence as a product of natural processes is a typical feature for traditional and vernacular housing settlements created by different users with their various needs. However, this feature disappears in housing production approaches for uniform and standard users due to mass production concerns. The main reason is that user needs cannot be fully transferred to design strategies due to excessive standardization. One of the main factors in this is the need for design tools and methods that effectively evaluate preferences with complex relationships. Shape grammars and spatial grammars, one of the generative design approaches, exhibit original emergence examples with their rich diversity and adaptability to different approaches. Hence, it can be seen as an alternative for solving such problems. This article presents a parametric spatial grammar model that can design high rise housing blocks with customized dwelling for each family. The compositions of these houses ground on specific spatial relationships defined by parametric rule sets. In the article, the generative design process was developed in two stages. The first stage includes the formal, syntactical, and functional relations of the spaces in a floor plan, their representation according to the rules, and generating alternative solutions. The second stage consists of the representation of the facade rules and the parametric grammar that generates the facades. In the developed spatial grammar model, user preferences are transferred to the computer through the designed interface. Dimensional properties of housing blocks are defined by a gridal system, which is determined according to user preferences and used as a starting base for spatial grammar. Plan layouts are generated on this defined grid and by sequential application of rules. In conclusion, the adaptability of spatial grammar and the rich alternatives designers can use and develop are emphasized.

Diffusion-based generative models have demonstrated remarkable capabilities in image synthesis, yet realistic hand generation remains a persistent challenge due to complex articulations, self-occlusion, and the lack of explicit structural guidance. To address these issues, we present SGMHand, a novel structure-guided hand inpainting framework that explicitly injects topological priors to enhance structural fidelity and spatial precision. Specifically, we present a structure-guided modulation (SGM) module that synergistically combines structure spatial attention with global feature calibration, enabling fine-grained geometric control over the generative process. Then, we devise a keypoint-aware (KA) loss that enforces topological coherence by aligning attention activations with structures, thereby bridging the gap between high-level semantics and low-level geometry. By jointly optimizing over structural constraints in both representation and learning objectives, SGMHand achieves semantically consistent and geometrically plausible hand synthesis, even under severe occlusion. Extensive experiments demonstrate the effectiveness and strong generalization ability of SGMHand across various foundation models, significantly enhancing the quality and realism of human image synthesis in diverse scenarios.

Presents a general and consistent design and fabrication process for pattern recognition LSI chips. In the process, the truth-table converted from the sample patterns is evolved to obtain generalization for unknown patterns, and the table is implemented on the reconfigurable LSI chips. With this process, named LoDETT (Logic Design with Evolving Truth-Table), the highly parallel hardware specialized for each task is easily fabricated with a quick turnaround time at each user site. We apply LoDETT to the English pronunciation recognition task and show the implementation results and performance.

A general direct analytical design process is presented for multiplexers having any number of channels with arbitrary channel complexity, bandwidths, and interchannel spacings. The theory assumes initially that independent doubly terminated designs are available for the individual filters, and formulas for modifications to parameters associated with the first two resonators are developed to match the multiplexer. These formulas are approximate, and the Iimitations of the theory are indicated with several computed examples. The theory is applied to the design of a five-channel interdigital multiplexer. A first-stage immittance compensation scheme is described which improves the design for limiting cases, but the theory of complete immittance compensation which handles even contiguous channel operation is reserved for a companion paper.