A novel feature-scrambling approach reveals the capacity of convolutional neural networks to learn spatial relations

Farahat, Amr; Effenberger, Felix; Vinck, Martin

doi:10.1016/j.neunet.2023.08.021

Cited by 4 publications

(2 citation statements)

References 47 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Buku ajar ini menggunakan model pembelajaran kooperatif Scramble, yang bertujuan untuk meningkatkan keterampilan sosial siswa dan meningkatkan keterlibatan dan hasil capaian akademik dan non akademik siswa, seperti kerja tim, komunikasi, dan pemecahan masalah (Yasin et al, 2020). Dengan menggunakan model pembelajaran Scramble, tujuan adalah untuk menarik perhatian siswa dan mendorong mereka untuk terlibat secara aktif dan mendorong mereka untuk berpikir kritis dan memahami konsep-konsep penting dengan memanfaatkan buku ajar (Farahat et al, 2023;Thelen, 2023). Beberapa keuntungan dari model pembelajaran Scramble termasuk evaluasi pemahaman siswa, umpan balik, dan mendorong mereka untuk berpikir kritis.…”

Section: Pendahuluanunclassified

Pengembangan Buku Ajar Ide Pokok Berbasis Potensi Lokal Pantai Tambakrejo Menggunakan Model Scramble Untuk Siswa Kelas IV SD

Nicky Nadila,

Sripit Widiastuti,

Adin Fauzi

2023

Madako Elementary School

View full text Add to dashboard Cite

Penelitian ini bertujuan untuk menjelaskan kelayakan Buku Ajar Berbasis Potensi Lokal Pantai Tambakrejo sebagai bahan mata pelajaran Bahasa Indonesia materi Ide Pokok. Metode penelitian dan pengembangan dengan tujuh tahapan: 1) Identifikasi potensi dan masalah; 2) Pengumpulan data; 3) Desain produk; 4) Validasi desain; 5) Revisi desain; 6) Pengujian produk; dan 7) Evaluasi produk. Validasi melibatkan ahli mata pelajaran, bahasa, dan media, serta uji keterbacaan oleh guru dan siswa melalui angket dan wawancara. Hasil validasi menunjukkan bahwa buku ini sangat layak, dengan persentase validasi ahli materi dan bahasa sebesar 91,66%, validasi ahli media 83,49%, serta uji keterbacaan oleh siswa 100% dan guru 92%. Dengan demikian dapat disimpulkan bahwa Buku Ajar ini layak digunakan sebagai materi ajar Bahasa Indonesia tentang Ide Pokok untuk siswa Kelas IV SD, dengan harapan dapat memudahkan pemahaman siswa dan pengajaran yang lebih kontekstual serta mempromosikan budaya lokal dalam pendidikan. Kata kunci: Buku Ajar, Potensi Lokal Tambakrejo, Model Kooperatif Scramble, Ide Pokok.

show abstract

Section: Pendahuluanunclassified

Pengembangan Buku Ajar Ide Pokok Berbasis Potensi Lokal Pantai Tambakrejo Menggunakan Model Scramble Untuk Siswa Kelas IV SD

Nicky Nadila,

Sripit Widiastuti,

Adin Fauzi

2023

Madako Elementary School

View full text Add to dashboard Cite

show abstract

“…Behaviourally, CNNs are prone to make errors with adversarial stimuli to which humans are resilient (Goodfellow et al, 2015;Nguyen et al, 2015), likely as a result of the divergence in representation deeper into the network (Feather et al, 2023). Recent works using images that were scrambled (Farahat et al, 2023) or distorted (Berardino et al, 2017) have found significant divergence in behaviour between humans and CNNs. In this study, we make a detailed comparison between human and CNN performance by constructing letter confusion matrices (similar to approaches used previously by Dodge and Karam, 2017;Borji and Itti, 2014).…”

Section: Introductionmentioning

confidence: 99%

Human visual performance for identifying letters affected by physiologically-inspired scrambling

Zhu,

Hess,

Baldwin

2024

Preprint

View full text Add to dashboard Cite

In human vision, the retinal input is transformed into internal representations through a series of stages. In earlier stages, the signals from a particular visual field locus are passed in parallel from one visual processing area to the next. The connections at each stage may therefore introduce "error", where incorrect or convergent projections result in a loss of spatial precision. Psychophysical and physiological studies have implicated spatial scrambling of this sort as a cause of the visual deficits in amblyopia. Several methods to measure scrambling (both in amblyopia and in healthy vision) have been developed in recent decades. In this work, we introduce a new approach. We consider two stages of visual processing where scrambling may occur: either at the input to or the output from the simple cell stage in V1. We refer to these as "subcortical" and "cortical" scrambling respectively. We investigated the impact of these two types of scrambling on a letter identification task. A physiologically-inspired decomposition and resynthesis algorithm was used to generate letter stimuli that simulate scrambling at each of these two stages. To establish a performance benchmark, we trained separate Convolutional Neural Networks (CNNs) to perform the task with each scrambling type. Comparing CNN performance against that of eight humans with normal healthy vision, we found humans exhibited greater resilience to subcortical scrambling compared to cortical scrambling. We further investigated performance by comparing confusion matrices. Compared to a simple template matching model, we found the human strategy to be more consistent with our CNNs. We conclude: i) the human resilience for subcortical scrambling suggests this may be the stage at which a greater degree of scrambling is introduced in the visual hierarchy, and ii) humans employ flexible strategies for identifying scrambled stimuli, more sophisticated than a simple template match to the expected target.

show abstract

DCDLN: A densely connected convolutional dynamic learning network for malaria disease diagnosis

Zhang,

Ding,

Zhang

et al. 2024

Neural Networks

View full text Add to dashboard Cite

A novel feature-scrambling approach reveals the capacity of convolutional neural networks to learn spatial relations

Cited by 4 publications

References 47 publications

Pengembangan Buku Ajar Ide Pokok Berbasis Potensi Lokal Pantai Tambakrejo Menggunakan Model Scramble Untuk Siswa Kelas IV SD

Pengembangan Buku Ajar Ide Pokok Berbasis Potensi Lokal Pantai Tambakrejo Menggunakan Model Scramble Untuk Siswa Kelas IV SD

Human visual performance for identifying letters affected by physiologically-inspired scrambling

DCDLN: A densely connected convolutional dynamic learning network for malaria disease diagnosis

Contact Info

Product

Resources

About