| Learning to draw: Emergent communication through sketching D Mihai, J Hare Advances in Neural Information Processing Systems 34, 7153-7166, 2021 | 22 | 2021 |
| Differentiable drawing and sketching D Mihai, J Hare arXiv preprint arXiv:2103.16194, 2021 | 22 | 2021 |
| Spatial and colour opponency in anatomically constrained deep networks E Harris, D Mihai, J Hare arXiv preprint arXiv:1910.11086, 2019 | 8 | 2019 |
| How convolutional neural network architecture biases learned opponency and color tuning E Harris, D Mihai, J Hare Neural Computation 33 (4), 858-898, 2021 | 5 | 2021 |
| Avoiding hashing and encouraging visual semantics in referential emergent language games D Mihai, J Hare arXiv preprint arXiv:1911.05546, 2019 | 4 | 2019 |
| The emergence of visual semantics through communication games D Mihai, J Hare arXiv preprint arXiv:2101.10253, 2021 | 2 | 2021 |
| Anatomically constrained ResNets exhibit opponent receptive fields; so what? EWA Harris, AD Mihai, J Hare | 2 | 2020 |
| Physically Embodied Deep Image Optimisation D Mihai, J Hare arXiv preprint arXiv:2201.08142, 2022 | | 2022 |
| Emergent visual communication AD Mihai University of Southampton, 2022 | | 2022 |
| Perceptions D Mihai, J Hare | | 2021 |
| Shared Visual Representations of Drawing for Communication: How do different biases affect human interpretability and intent? D Mihai, J Hare arXiv preprint arXiv:2110.08203, 2021 | | 2021 |
| Understanding the Meaning of Images D Mihai, J Hare | | |
Jonathon HareProfessor of Machine Learning and Computer Vision, University of Southampton, United KingdomVerified email at ecs.soton.ac.uk
