Differentiable Learning Research Articles

Penggunaan Media Berbasis Lingkungan Dalam Meningkatkan Kreativitas Anak Usia Dini

This study was motivated by the lack of variety in early childhood education learning activities and a lack of inclination toward ideas, creativity, or toughness. The purpose of this study is to examine the effect of using environment – based learning media on students’ creativity in early childhood. The problem was devised about how the use of environment-based learning media increases the creativity of early childhood education. This study used a quantitative research design. The result of this study indicated the improvement of early childhood creativity by using environment-based learning media. The result showed that the t-test is 3.294 and the t-table is 1.796. This differentiable learning was used by providing game variation, stimulating the children to develop and try to choose activities accordance with their study style, and also getting the real experience of using the learning media from the surrounding environment. So that the children’s exploration and their imagination have the opportunity to develop creativity and critical thinking.

A Noise-Tolerant Differentiable Learning Approach for Single Occurrence Regular Expression with Interleaving

We study the problem of learning a single occurrence regular expression with interleaving (SOIRE) from a set of text strings possibly with noise. SOIRE fully supports interleaving and covers a large portion of regular expressions used in practice. Learning SOIREs is challenging because it requires heavy computation and text strings usually contain noise in practice. Most of the previous studies only learn restricted SOIREs and are not robust on noisy data. To tackle these issues, we propose a noise-tolerant differentiable learning approach SOIREDL for SOIRE. We design a neural network to simulate SOIRE matching and theoretically prove that certain assignments of the set of parameters learnt by the neural network, called faithful encodings, are one-to-one corresponding to SOIREs for a bounded size. Based on this correspondence, we interpret the target SOIRE from an assignment of the set of parameters of the neural network by exploring the nearest faithful encodings. Experimental results show that SOIREDL outperforms the state-of-the-art approaches, especially on noisy data.

Differentiable learning of matricized DNFs and its application to Boolean networks

Boolean networks (BNs) are well-studied models of genomic regulation in biology where nodes are genes and their state transition is controlled by Boolean functions. We propose to learn Boolean functions as Boolean formulas in disjunctive normal form (DNFs) by an explainable neural network Mat_DNF and apply it to learning BNs. Directly expressing DNFs as a pair of binary matrices, we learn them using a single layer NN by minimizing a logically inspired non-negative cost function to zero. As a result, every parameter in the network has a clear meaning of representing a conjunction or literal in the learned DNF. Also we can prove that learning DNFs by the proposed approach is equivalent to inferring interpolants in logic between the positive and negative data. We applied our approach to learning three literature-curated BNs and confirmed its effectiveness. We also examine how generalization occurs when learning data is scarce. In doing so, we introduce two new operations that can improve accuracy, or equivalently generalizability for scarce data. The first one is to append a noise vector to the input learning vector. The second one is to continue learning even after learning error becomes zero. The first one is explainable by the second one. These two operations help us choose a learnable DNF, i.e., a root of the cost function, to achieve high generalizability.

Differentiable learning of rules with constants in knowledge graph

Knowledge reasoning helps overcome the incompleteness of knowledge graphs (KGs) and has significantly contributed to the development of large KGs. Rule mining, one of the key tasks of knowledge reasoning, studies the problem of learning interpretable inference patterns over KGs. Existing rule mining methods mainly focus on learning rules that consist of different relations and variables, restricting the form of rules to be closed path. While rules could be diverse and in order to enrich the forms of rules, we argue that constants should also be considered in the rule mining process. We propose an Elegant Differentiable rUle learning with Constant mEthod (EduCe).11Source code of EduCeis available at https://github.com/yep96/Educe. We propose a constant operator and dynamic weight mechanism, which choose the constants that should be added and decrease the number of parameters, respectively. The model could mine diverse and accurate rules in an efficient way with these modules. The experimental results on several knowledge graph completion benchmarks show that EduCe achieves state-of-the-art link prediction results among differentiable rule mining methods and successfully learns diverse and high-quality rules.

Differentiable Learning of Scalable Multi-Agent Navigation Policies

We present an end-to-end differentiable learning algorithm for multi-agent navigation policies. Compared with prior model-free learning algorithms, our method leads to a significant speedup via the gradient information. Our key innovation lies in a novel differentiability analysis of the optimization-based crowd simulation algorithm via the implicit function theorem. Inspired by continuum multi-agent modeling techniques, we further propose a kernel-based policy parameterization, allowing our learned policy to scale up to an arbitrary number of agents without re-training. We evaluate our algorithm on two tasks in obstacle-rich environments, partially labeled navigation and evacuation, for which loss functions can be defined making the entire task learnable in an end-to-end manner. The results show that our method can achieve more than one order of magnitude speedup over model-free baselines and readily scale to unseen target configurations and agent sizes.

Differentiable Learning of Sequence-Specific Minimizer Schemes with DeepMinimizer.

Minimizers are widely used to sample representative k-mers from biological sequences in many applications, such as read mapping and taxonomy prediction. In most scenarios, having the minimizer scheme select as few k-mer positions as possible (i.e., having a low density) is desirable to reduce computation and memory cost. Despite the growing interest in minimizers, learning an effective scheme with optimal density is still an open question, as it requires solving an apparently challenging discrete optimization problem on the permutation space of k-mer orderings. Most existing schemes are designed to work well in expectation over random sequences, which have limited applicability to many practical tools. On the other hand, several methods have been proposed to construct minimizer schemes for a specific target sequence. These methods, however, only approximate the original objective with likewise discrete surrogate tasks that are not able to significantly improve the density performance. This article introduces the first continuous relaxation of the density minimizing objective, DeepMinimizer, which employs a novel Deep Learning twin architecture to simultaneously ensure both validity and performance of the minimizer scheme. Our surrogate objective is fully differentiable and, therefore, amenable to efficient gradient-based optimization using GPU computing. Finally, we demonstrate that DeepMinimizer discovers minimizer schemes that significantly outperform state-of-the-art constructions on human genomic sequences.

Deep differentiable reinforcement learning and optimal trading

In many reinforcement learning applications, the underlying environment reward and transition functions are explicitly known differentiable functions. This enables us to use recent research which applies machine learning tools to stochastic control to find optimal action functions. In this paper, we define differentiable reinforcement learning as a particular case of this research. We find that incorporating deep learning in this framework leads to more accurate and stable solutions than those obtained from more generic actor critic algorithms. We apply this deep differentiable reinforcement learning (DDRL) algorithm to the problem of one asset optimal trading strategies in various environments where the market dynamics are known. Thanks to the stability of this method, we are able to efficiently find optimal strategies for complex multi-scale market models. We also extend these methods to simultaneously find optimal action functions for a wide range of environment parameters. This makes it applicable to real life financial signals and portfolio optimization where the expected return has multiple time scales. In the case of a slow and a fast alpha signal, we find that the optimal trading strategy consists in using the fast signal to time the trades associated to the slow signal.

Neural nano-optics for high-quality thin lens imaging

Nano-optic imagers that modulate light at sub-wavelength scales could enable new applications in diverse domains ranging from robotics to medicine. Although metasurface optics offer a path to such ultra-small imagers, existing methods have achieved image quality far worse than bulky refractive alternatives, fundamentally limited by aberrations at large apertures and low f-numbers. In this work, we close this performance gap by introducing a neural nano-optics imager. We devise a fully differentiable learning framework that learns a metasurface physical structure in conjunction with a neural feature-based image reconstruction algorithm. Experimentally validating the proposed method, we achieve an order of magnitude lower reconstruction error than existing approaches. As such, we present a high-quality, nano-optic imager that combines the widest field-of-view for full-color metasurface operation while simultaneously achieving the largest demonstrated aperture of 0.5 mm at an f-number of 2.

Analyzing Differentiable Fuzzy Logic Operators

The AI community is increasingly putting its attention towards combining symbolic and neural approaches, as it is often argued that the strengths and weaknesses of these approaches are complementary. One recent trend in the literature is weakly supervised learning techniques that employ operators from fuzzy logics. In particular, these use prior background knowledge described in such logics to help the training of a neural network from unlabeled and noisy data. By interpreting logical symbols using neural networks, this background knowledge can be added to regular loss functions, hence making reasoning a part of learning.We study, both formally and empirically, how a large collection of logical operators from the fuzzy logic literature behave in a differentiable learning setting. We find that many of these operators, including some of the most well-known, are highly unsuitable in this setting. A further finding concerns the treatment of implication in these fuzzy logics, and shows a strong imbalance between gradients driven by the antecedent and the consequent of the implication. Furthermore, we introduce a new family of fuzzy implications (called sigmoidal implications) to tackle this phenomenon. Finally, we empirically show that it is possible to use Differentiable Fuzzy Logics for semi-supervised learning, and compare how different operators behave in practice. We find that, to achieve the largest performance improvement over a supervised baseline, we have to resort to non-standard combinations of logical operators which perform well in learning, but no longer satisfy the usual logical laws.

Non‐Differentiable Leaning of Quantum Circuit Born Machine with Genetic Algorithm

The Quantum Circuit Born Machine (QCBM) is a generative quantum machine learning model that can be efficiently trained and run on the NISQ-era quantum processors. It is likely that QCBM will be one of the first quantum machine learning models to find productive applications in quantitative finance as a powerful market generator. In this paper we test QCBM performance on a dataset of spot FX log-returns (heavy-tailed distribution) as well as specially constructed mixture of Normal distributions (which models spiky light-tailed distribution). The QCBM has greater expressive power than comparable classical neural networks such as restricted Boltzmann machine (RBM) and, therefore, has potential to demonstrate quantum advantage by generating high-quality samples from the learned empirical distribution of the market risk factors while using less computational resources than its classical counterpart. However, efficient training of QCBM remains a challenging problem. Traditional differentiable learning approach may not work well when the loss function is highly non-smooth. In such cases it may be more efficient to use the non-differentiable learning methods. This paper proposes a non-differentiable learning approach to the training of QCBM based on genetic glgorithm (GA). The paper also presents results of the numerical experiments which compare performance of QCBM trained with GA against performance of the equivalent classical RBM and investigates the question of GA convergence as a function of QCBM architecture and the choice of algorithm's hyperparameters.

End-to-End Differentiable Learning to HDR Image Synthesis for Multi-exposure Images

Recently, high dynamic range (HDR) image reconstruction based on the multiple exposure stack from a given single exposure utilizes a deep learning framework to generate high-quality HDR images. These conventional networks focus on the exposure transfer task to reconstruct the multi-exposure stack. Therefore, they often fail to fuse the multi-exposure stack into a perceptually pleasant HDR image as the inversion artifacts occur. We tackle the problem in stack reconstruction-based methods by proposing a novel framework with a fully differentiable high dynamic range imaging (HDRI) process. By explicitly using the loss, which compares the network's output with the ground truth HDR image, our framework enables a neural network that generates the multiple exposure stack for HDRI to train stably. In other words, our differentiable HDR synthesis layer helps the deep neural network to train to create multi-exposure stacks while reflecting the precise correlations between multi-exposure images in the HDRI process. In addition, our network uses the image decomposition and the recursive process to facilitate the exposure transfer task and to adaptively respond to recursion frequency. The experimental results show that the proposed network outperforms the state-of-the-art quantitative and qualitative results in terms of both the exposure transfer tasks and the whole HDRI process.

End-to-end differentiable learning of turbulence models from indirect observations

The emerging push of the differentiable programming paradigm in scientific computing is conducive to training deep learning turbulence models using indirect observations. This paper demonstrates the viability of this approach and presents an end-to-end differentiable framework for training deep neural networks to learn eddy viscosity models from indirect observations derived from the velocity and pressure fields. The framework consists of a Reynolds-averaged Navier–Stokes (RANS) solver and a neural-network-represented turbulence model, each accompanied by its derivative computations. For computing the sensitivities of the indirect observations to the Reynolds stress field, we use the continuous adjoint equations for the RANS equations, while the gradient of the neural network is obtained via its built-in automatic differentiation capability. We demonstrate the ability of this approach to learn the true underlying turbulence closure when one exists by training models using synthetic velocity data from linear and nonlinear closures. We also train a linear eddy viscosity model using synthetic velocity measurements from direct numerical simulations of the Navier–Stokes equations for which no true underlying linear closure exists. The trained deep-neural-network turbulence model showed predictive capability on similar flows.

Guest Editorial Introduction to the Special Section on Intelligent Visual Content Analysis and Understanding

Visual content analysis and understanding attract tremendous attention because of its potentially wide range of applications including human activity analysis, automated photo face tagging, multicamera tracking, crowded counting, and biometric security. With recent progress in end-to-end differentiable learning, the accuracy of algorithms has been significantly improved and even outperforms humans in some tasks. In addition, multimodality methods, targeting on making full use of various visual data sources, are further investigated. These developments contribute to the innovations of two core modules for a typical intelligent vision system, i.e., image and video description and recognition, which are critical for the success of the visual content analysis and understanding in more complex and challenging open world.

Using Deep Learning to Predict Fracture Patterns in Crystalline Solids

Summary Fracture is a catastrophic process whose understanding is critical for evaluating the integrity and sustainability of engineering materials. Here, we present a machine-learning approach to predict fracture processes connecting molecular simulation into a physics-based data-driven multiscale model. Based on atomistic modeling and a novel image-processing approach, we compile a comprehensive training dataset featuring fracture patterns and toughness values for different crystal orientations. Assessments of the predictive power of the machine-learning model shows excellent agreement not only regarding the computed fracture patterns but also the fracture toughness values and is examined for both mode I and mode II loading conditions. We further examine the ability of predicting fracture patterns in bicrystalline materials and material with gradients of microstructural crystal orientation. These results further underscore the excellent predictive power of our model. Potential applications of this model could be widely applied in material design.

Non-Differentiable Learning of Quantum Circuit Born Machine with Genetic Algorithm

The Quantum Circuit Born Machine (QCBM) is a generative quantum machine learning model that can be efficiently trained and run on the NISQ era quantum processors. QCBM has greater expressive power than comparable classical neural networks such as Restricted Boltzmann Machine (RBM) and, therefore, has potential to demonstrate quantum advantage by generating high quality samples from the learned empirical distribution while using less computational resources than its classical counterpart. However, efficient training of QCBM remains a challenging problem. Traditional differentiable learning approach may not work well when the loss function is highly non-smooth. In such cases it may be more efficient to use the non-differentiable learning methods. This paper proposes a non-differentiable learning approach to the training of QCBM based on Genetic Algorithm (GA). The paper also presents results of the numerical experiments which compare performance of QCBM trained with GA against performance of the equivalent classical RBM and investigates the question of GA convergence as a function of QCBM architecture and the choice of algorithm’s hyperparameters.

When Does Communication Learning Need Hierarchical Multi-Agent Deep Reinforcement Learning

Multi-agent systems need to communicate to coordinate a shared task. We show that a recurrent neural network (RNN) can learn a communication protocol for coordination, even if the actions to coordinate are performed steps after the communication phase. We show that a separation of tasks with different temporal scale is necessary for successful learning. We contribute a hierarchical deep reinforcement learning model for multi-agent systems that separates the communication and coordination task from the action picking through a hierarchical policy. We further on show, that a separation of concerns in communication is beneficial but not necessary. As a testbed, we propose the Dungeon Lever Game and we extend the Differentiable Inter-Agent Learning (DIAL) framework. We present and compare results from different model variations on the Dungeon Lever Game.

Learning and inference on generative adversarial quantum circuits

Quantum mechanics is inherently probabilistic in light of Born's rule. Using quantum circuits as probabilistic generative models for classical data exploits their superior expressibility and efficient direct sampling ability. However, training of quantum circuits can be more challenging compared to classical neural networks due to lack of efficient differentiable learning algorithm. We devise an adversarial quantum-classical hybrid training scheme via coupling a quantum circuit generator and a classical neural network discriminator together. After training, the quantum circuit generative model can infer missing data with quadratic speed up via amplitude amplification. We numerically simulate the learning and inference of generative adversarial quantum circuit using the prototypical Bars-and-Stripes dataset. Generative adversarial quantum circuits is a fresh approach to machine learning which may enjoy the practically useful quantum advantage on near-term quantum devices.

End-to-End Differentiable Learning of Protein Structure.

Predicting protein structure from sequence is a central challenge of biochemistry. Co-evolution methods show promise, but an explicit sequence-to-structure map remains elusive. Advances in deep learning that replace complex, human-designed pipelines with differentiable models optimized end to end suggest the potential benefits of similarly reformulating structure prediction. Here, we introduce an end-to-end differentiable model for protein structure learning. The model couples local and global protein structure via geometric units that optimize global geometry without violating local covalent chemistry. We test our model using two challenging tasks: predicting novel folds without co-evolutionary data and predicting known folds without structural templates. In the first task, the model achieves state-of-the-art accuracy, and in the second, it comes within 1-2Å; competing methods using co-evolution and experimental templates have been refined over many years, and it is likely that the differentiable approach has substantial room for further improvement, with applications ranging from drug discovery to protein design.

Differentiable learning of quantum circuit Born machines

Quantum circuit Born machines are generative models which represent the probability distribution of classical dataset as quantum pure states. Computational complexity considerations of the quantum sampling problem suggest that the quantum circuits exhibit stronger expressibility compared to classical neural networks. One can efficiently draw samples from the quantum circuits via projective measurements on qubits. However, similar to the leading implicit generative models in deep learning, such as the generative adversarial networks, the quantum circuits cannot provide the likelihood of the generated samples, which poses a challenge to the training. We devise an efficient gradient-based learning algorithm for the quantum circuit Born machine by minimizing the kerneled maximum mean discrepancy loss. We simulated generative modeling of the Bars-and-Stripes dataset and Gaussian mixture distributions using deep quantum circuits. Our experiments show the importance of circuit depth and gradient-based optimization algorithm. The proposed learning algorithm is runnable on near-term quantum device and can exhibit quantum advantages for generative modeling.

Effect of Problem Based Learning through Differentiable Learning Styles in the Teaching of Statistics

The goal of this study was to evaluate the effect of Problem Based Learning (PBL) using a strategy mediated by learning styles on variables of academic performance, attitude and motivation of students within a Statistics online course. Participants in the study were alumni of Statistics II subject course at the University of Guadalajara’s Regional Campus in Autlan. Two groups were formed: an experimental one lead through online supported instructional PBL strategy, and a control one attended in a classroom (traditional way). Three questionnaires were applied: first one identified predominant learning styles of students (PEPS), the second one attitude (CUAC) and third one was to interpret motivation (CUMO) towards the proposed learning environment; which were applied at the beginning and at the end of the course. Results of T-test and ANOVA test showed that the learning styles are not related to the instructional strategy of PBL; no differences were found in variables, attitude nor motivation when learning styles were compared within the PBL group. However, such results showed differences in academic performance, attitude and motivation among students of the PBL online group compared to the one that participated in a traditional way.