233 research outputs found
MDP homomorphic networks: Group symmetries in reinforcement learning
This paper introduces MDP homomorphic networks for deep reinforcement learning. MDP homomorphic networks are neural networks that are equivariant under symmetries in the joint state-action space of an MDP. Current approaches to deep reinforcement learning do not usually exploit knowledge about such structure. By building this prior knowledge into policy and value networks using an equivariance constraint, we can reduce the size of the solution space. We specifically focus on group-structured symmetries (invertible transformations). Additionally, we introduce an easy method for constructing equivariant network layers numerically, so the system designer need not solve the constraints by hand, as is typically done. We construct MDP homomorphic MLPs and CNNs that are equivariant under either a group of reflections or rotations. We show that such networks converge faster than unstructured baselines on CartPole, a grid world and Pong.Interactive Intelligenc
Plannable Approximations to MDP Homomorphisms: Equivariance under Actions
This work exploits action equivariance for representation learning in reinforcement learning. Equivariance under actions states that transitions in the input space are mirrored by equivalent transitions in latent space, while the map and transition functions should also commute. We introduce a contrastive loss function that enforces action equivariance on the learned representations. We prove that when our loss is zero, we have a homomorphism of a deterministic Markov Decision Process (MDP). Learning equivariant maps leads to structured latent spaces, allowing us to build a model on which we plan through value iteration. We show experimentally that for deterministic MDPs, the optimal policy in the abstract MDP can be successfully lifted to the original MDP. Moreover, the approach easily adapts to changes in the goal states. Empirically, we show that in such MDPs, we obtain better representations in fewer epochs compared to representation learning approaches using reconstructions, while generalizing better to new goals than model-free approaches.Interactive Intelligenc
SISTEM INFORMASI PEMESANAN BARANG PADA CV RESTU PERDANA PALEMBANG BERBASIS WEB
Pesatnya perkembangan komputer dan teknologi informasi menjadi pemicu berbagai perusahaan untuk memanfaatkan sistem informasi berbasis dekstop maupun Web. CV. Restu Perdana Palembang pun membutuhkan sebuah sistem informasi untuk memperoleh peluang bersaing dan meningkatkan eksistensi perusahaan. Permasalahan yang dialami oleh CV. Restu Perdana Palembang yaitu belum adanya media pemasaran mengakibatkan sulitny perusahaan memasarkan produk maupun meningkatkan produktifitas, kurang cepat merespon keluhan pelanggan, dan kebutuhan pelanggan akan transaksi kapan saja dan di mana saja. Metodologi yang digunakan adalah RUP (Rational Unified Process) dengan metode iterasi. Implementasi desain program menggunakan bahasa PHP6, Sedangkan DBMS menggunakan MySQL. Dalam sistem informasi ini terdapat fungsi untuk Login dalam 2 versi yaitu pelanggan tetap dan pelanggan biasa ,register ,ubah password , pesan produk , tambah jumlah pesa produk ,cari produk input ,pembayaran ,tambah data admin ,tambah data produk ,input berita , input data produk ,input kritik dan saran . Dengan adanya sistem informasi pemesanan berbasis web ini dapat memberikan kemudahan dan meminimalkan kesalahan yang terjadi pada bagian administrasi serta stakeholders yang terkait dengan kegiatan pemesanan tersebu
APLIKASI SISTEM AKADEMIK PADA SEKOLAH MENENGAH PERTAMA NEGERI 55 PALEMBANG
Tujuan penulisan tugas akhir ini dilakukan untuk membangun sistem
informasi akademik pada sekolah menengah pertama negeri 55 Palembang. Pada
sistem yang berjalan saat ini berdasarkan hasil pengamatan yang dilakukan penulis
ditemukan beberapa kendala seperti masih lambatnya pencarian data, dikarenakan
penyimpanan data masih dalam bentuk fisik atau berupa dokumen. Disamping itu
seringnya terjadi kesalahan pencatatan data, laporan-laporan yang mengalami
keterlambatan dan kurang efisiensinya sistem yang digunakan, menjadi suatu kendala
yang harus diatasi. Metode yang digunakan dalam penyusunan tugas akhir ini adalah
metode Iteration dan juga menggunakan metode pengumpulan data yang terdiri dari
beberapa metode antara lain metode wawancara, metode observasi dan kepustakaan.
Hasil yang dicapai adalah suatu aplikasi sistem manajemen yang berbasis komputer
yang dapat memudahkan dalam menyelesaikan permasalahan dan hambatan dalam
kegiatan akademik. Kesimpulan tugas akhir ini adalah suatu sistem informasi yang
berbasis komputer yang dapat mengatasi permasalahan-permasalahan pencarian data,
penyimpanan data, pembuatan laporan dan efisiensi data dibidang akademik
Tc-99m MDP Uptake in Retroperitoneal Fibrosis
A 60-year-old man with a history of surgical resection of leiomyosarcoma of the right kidney had a bone scan to evaluate possible, metastatic osseous disease. The bone scan showed significantly increased Tc-99m MDP activity in the right abdomen, corresponding to a heterogeneous mass in the right retroperitoneal region on the subsequent abdominal CT scan. These findings were initially interpreted as recurrent malignancy. However, the surgical pathology examination demonstrated that the lesion was retroperitoneal fibrosis, which was not previously known to have increased MDP uptake.Radiology, Nuclear Medicine & Medical ImagingSCI(E)PubMed6EDITORIAL MATERIAL3189-1903
PEMBANGUNAN APLIKASI PEMBELAJARAN ANATOMI TUBUH MANUSIA BERBASIS MULTIMEDIA
Human anatomy is one of the subject matter of biology which deals with human body
parts. Students need a picture to explain the material object. Therefore, should made
in a medium of learning that makes the material easy to understand, such as by using
the facilities that already exist by using computer technology. This thesis aims to
overcome this problem, ranging from providing visual object that is not boring in
teaching and learning through the interactive nature of which is given to teachers can
easily convey the material. The author uses Macromedia Flash 8 media in making the
application. In addition, the system development methodology that author use is
Waterfall methodology. Applications are also tested and evaluated using
questionnaires. Based on the questionnaire, the respondents agreed that the
application meets user needs. The author hopes that the application can become an
alternative media in the learning process for students and teachers can find a solution
to overcomethe problem of learning in the classroom by using the advanced and
innovative media in accordance with current technological developments
MOBILE MAP EXPLORER LOKASI BENGKEL DI PALEMBANG DENGAN GPS BERBASIS PLATFORM ANDROID
Mobile Map Explorer application of the workshop location is Android based application
that are used to locate the nearest of workshop using GPS. The existence of the workshop
is required by automotive user to take care or repair the vehicle. Until now, that has been a
lot of digital maps are available, but not all locations can be displayed, particularly the
location of the workshop in Palembang. Based of these problems, the author developed a
mobile application about map explorer of workshop location with the aim to help
automotive user to locate the nearest the workshop in Palembang on the phone by using the
Global Positioning System (GPS) shown on the map. This application is built on the
Android platform, and therefore the mobile phones that support this application is a phone
based on Android because this application will be installed first on a mobile phone users.
The research methodology used by author is method of Rational Unified Process (RUP),
which consists of four stages, namely Inception phase, Elaboration, Construction, and
Transition. Evaluation to do by distributing questionnaires. With this application, the
author hope to help users in searching and finding the location of the nearest the workshop
to take care or repair the vehicle
An application of the ABS LX algorithm to multiple sequence alignement
Il problema del multiple sequence alignement porta alla soluzione di sistemi lineari malcondizionati, per il cui trattamento si è usato con esiti positivi il metodo implicito LOX della classe ABSWe present an application of ABS algorithms for multiple sequence alignment (MSA). The Markov decision process (MDP) based model leads to a linear programming problem (LPP), whose solution is linked to a suggested alignment. The important features of our work include the facility of alignment of multiple sequences simultaneously and no limit for the length of the sequences. Our goal here is to avoid the excessive computing time, needed by dynamic programming based algorithms for alignment of a large number of sequences. In an attempt to demonstrate the integration of the ABS approach with complex mathematical frameworks, we apply the ABS implicit LX algorithm to elucidate the LPP, constructed with the assistance of MDP. The MDP applied for MSA is a pragmatic approach and entails a scope for future work. Programming is done in the MATLAB environment
RANCANG BANGUN SISTEM INFORMASI PERJALANAN IBADAH UMRAH BERBASIS WEB PADA PT. BINA INSANI MADINAH PALEMBANG
PT. BinaInsaniMadinah Palembang is a company engaged in Umrah Tour Services. PT. BinaInsaniMadinah Palembang was encouraged to become a company that promoted the best and professional services to all worshipers. For its business processes, PT. BinaInsaniMadinah Palembang has used a computer by using Microsoft Office application as a media to manage data for pilgrims who order Umrah packages. The problems faced so far are information on umrah pilgrimage sometimes changing so that the lack of pilgrims in getting information on Umrah pilgrimages, the registration process takes a long time because prospective pilgrims must come to the office to fill in the registration form sometimes prospective pilgrims fill out the wrong forms so candidates pilgrims have to fill out a new registration form, it is difficult for companies to recapitulate data reports because the pilgrims' data reports still use archives. Then the information system design is proposed as a solution to these constraints. In the process of development, the author uses an iterative research methodology. Later in the development of the application using notepad ++ as an editor with the language of the PHP program and MySQL as the data storage media. Information about promotions, departure schedules, rituals, activity schedules and online registration are the outputs produced by this applicatio
RANCANG BANGUN APLIKASI PEMBELAJARAN PERANGKAT LAPTOP BERBASIS AUGMENTED REALITY
This e-Learning application uses Augmented Reality technology. Augmented Reality
is the interaction of cyberspace and real. Applications developed with ActionScript
language. Making 3D objects using 3ds Max 2010 and the application is implemented
in web applications. The object used in the Augmented Reality applications is
Notebook devices. Markers that are used not only one marker but a lot of markers. In
making this application, the author used Prototype methodology
- …
