Not only have deep networks become standard in machine learning, they are increasingly of interest in neuroscience as models of cortical computation that capture relationships between structural and functional properties. In addition they are a useful target of theoretical research into the properties of network computation. Deep networks typically have a serial or approximately serial organization across layers, and this is often mirrored in models that purport to represent computation in mammalian brains. There are, however, multiple examples of parallel pathways in mammalian brains. In some cases, such as the mouse, the entire visual system appears arranged in a largely parallel, rather than serial fashion. While these pathways may be formed by differing cost functions that drive different computations, here we present a new mathematical analysis of learning dynamics in networks that have parallel computational pathways driven by the same cost function. We use the approximation of deep linear networks with large hidden layer sizes to show that, as the depth of the parallel pathways increases, different features of the training set (defined by the singular values of the input-output correlation) will typically concentrate in one of the pathways. This result is derived analytically and demonstrated with numerical simulation. Thus, rather than sharing stimulus and task features across multiple pathways, parallel network architectures learn to produce sharply diversified representations with specialized and specific pathways, a mechanism which may hold important consequences for codes in both biological and artificial systems.

Not only have deep networks become standard in machine learning, they are increasingly of interest in neuroscience as models of cortical computation that capture relationships between structural and functional properties. In addition they are a useful target of theoretical research into the properties of network computation. Deep networks typically have a serial or approximately serial organization across layers, and this is often mirrored in models that purport to represent computation in mammalian brains. There are, however, multiple examples of parallel pathways in mammalian brains. In some cases, such as the mouse, the entire visual system appears arranged in a largely parallel, rather than serial fashion.

The recent developments in soft computing cannot be complete without noting the contributions of artificial neural machine learning systems that draw inspiration from real cortical tissue or processes that occur in human brain. The universal approximability of such neural systems has led to its wide spread use, and novel developments in this evolving technology has shown that there is a bright future for such Artificial Intelligent (AI) techniques in the soft computing field. Indeed, the proliferation of large and very deep networks of artificial neural systems and the corresponding enhancement and development of neural machine learning algorithms have contributed immensely to the development of the modern field of Deep Learning as may be found in the well documented research works of Lecun, Bengio and Hinton. However, the key requirements of end user affordability in addition to reduced complexity and reduced data learning size requirement means there still remains a need for the synthesis of more cost-efficient and less data-hungry artificial neural systems. In this report, we present an overview of a new competing bio-inspired continual learning neural tool Neuronal Auditory Machine Intelligence (Neuro-AMI) as a predictor detailing its functional and structural details, important aspects on right applicability, some recent application use cases and future research directions for current and prospective machine learning experts and data scientists.

How does the brain represent the world and allow spatial navigation? One mechanism is hippocampal place cells—neurons that fire according to where an animal is in its environment. Different place cells fire according to different locations, and together they are thought to provide a cognitive map that supports spatial navigation and memory ([ 1 ][1]). Place cells have been described in a range of mammalian species, including mice, bats, marmosets, and humans. However, most studies have used rats in small enclosures or mazes. Thus, it is unknown how such representations might underpin larger-scale, real-world navigation. On page 933 of this issue, Eliav et al. ([ 2 ][2]) show that in bats flying in a large (200-m-long) enclosure, most place cells fire in several different locations and with varying spatial scales. Such multiscale representations are likely the most efficient way for a finite number of neurons to encode large distances. Neurophysiological recordings in rats exploring relatively small “open-field” environments (∼1 m2) or running along short tracks 1 to 2 m long have revealed that a given place cell in the hippocampus typically fires when the rat is in a single area within the apparatus (called its place field) ([ 1 ][1], [ 3 ][3], [ 4 ][4]). In the few experiments that have investigated bigger open-field environments and longer tracks, place fields are typically slightly enlarged compared with those in smaller environments ([ 4 ][4]–[ 6 ][5]), and individual place cells in CA1 (the main output region of the hippocampus) fire in multiple, irregularly spaced locations ([ 5 ][6], [ 6 ][5]), with more place fields per cell in tracks of increasing length ([ 6 ][5]). Within a given environment, the different place fields of each hippocampal neuron are of a fairly uniform size, but there is an anatomical gradient, with the most dorsal hippocampal place cells having the smallest fields and ventral hippocampal cells having the largest fields ([ 3 ][3], [ 7 ][7]). Together, these studies suggest that the hippocampus provides an ensemble place code, whereby different combinations of neurons are active in any given location, and that coding of different spatial scales is provided by different neurons across the dorsal-ventral hippocampal axis. But how does the mammalian brain represent much larger spaces, on the spatial scale that animals would need to navigate in their natural environment? Eliav et al. wirelessly recorded from dorsal CA1 place cells in bats as they flew along a 200-m-long tunnel between two feeding stations. They found not only that place cells expressed multiple, irregularly spaced place fields in this very large environment but also that the size of the different place fields expressed by a given neuron varied widely: The mean ratio of the largest:smallest field was 4.4:1, but this was as high as 20:1 in some cells (see the figure). By contrast, and consistent with observations in rats, in a shorter 6-m-long tunnel, place cells expressed only one or two fields, the average field size was smaller than in the 200-mlong tunnel, and fields of the same cell were of a similar size (mean ratio <2:1). ![Figure][8] Navigating large, complex spaces Eliav et al. found that bats exhibit multiscale place cell coding. Individual place cells in the hippocampus fire according to a range of spatial scales (place fields of a single place cell indicated by circles), allowing optimal processing of a large environment with a finite number of cells. GRAPHIC: N. DESAI/ SCIENCE These findings of multiscale coding by individual place cells may help answer a puzzling question: How can a finite population of place cells encode the large environments in which mammals navigate in the wild, at both large and small spatial scales? The modeling by Eliav et al. shows that the multiscale coding mechanism seen in the bats is a particularly efficient mechanism for coding large environments. It needs fewer neurons for accurate decoding of the current location of the bat than other ensemble coding mechanisms based on individual cells having multiple fields of the same size and other cells having fields of different sizes (as had previously been assumed). It will be important to determine the extent to which multiscale coding by individual neurons is a general property of hippocampal coding across species and across different types and scales of environments. A preliminary study of rats following a moving robotic feeder in an 18.6-m2 open-field environment reported that cells in dorsal CA1 exhibited the same type of multiscale coding as found in the tunnel-flying bats ([ 8 ][9]). This indicates that this type of firing may be a general principle of hippocampal coding of large-scale space across mammalian species. Moreover, perhaps in large, continuous spaces, multiscale place cell representation may be the rule. As with many elegant studies, the work of Eliav et al. points to promising new avenues of research. One key question is how multiscale encoding arises. The two main inputs to CA1 (where the multiscale place cells have been described) are the CA3 and the medial entorhinal cortex (MEC). CA3 also contains place cells; indeed, the dorsal-ventral gradient of small-large place fields was described in CA3 neurons in rats ([ 7 ][7]). Conversely, the MEC contains a different type of spatial cell called grid cells. Each grid cell fires in multiple locations arranged in a regular hexagonal grid pattern that repeats across the environment (again with a dorsal-ventral arrangement of grid field size and spacing) ([ 9 ][10], [ 10 ][11]). Grid cells are thought to be important for path integration, where animals use self-motion signals to estimate distances and directions traveled. Eliav et al. suggest a feed-forward model whereby the multiscale fields in CA1 result from convergence of inputs from multiple CA3 place cells with different spatial scales onto each CA1 place cell. Predictions of this model that still need to be tested are that CA3 neurons should not show multiple fields in large environments and that either grid cells should not show multiple fields or grid cell inputs do not contribute to the firing of CA1 place fields in large environments. A second question is whether there is a continuum of multiscale coding across environments of all sizes or whether (as suggested by Eliav et al. ) multiscale coding occurs only in sufficiently large environments. And if the latter, what behavioral, perceptual, and neural mechanisms trigger the transition from small-scale to large-scale encoding of space? The study of Eliav et al. provides a marker for the need to examine spatial coding in ethologically relevant environments. The multiscale place cell coding mechanism that they demonstrate may allow both fine-scale spatial localization and localization on a more extended scale, which would be required for navigating accurately between very distant locations hundreds of meters or kilometers apart. It will be interesting to see whether similar multiscale spatial representations occur in humans or nonhuman primates navigating (virtual or real) large, open spaces and whether multiscale coding by individual neurons occurs in other, nonspatial domains, such as the coding of time ([ 11 ][12]). 1. [↵][13]1. J. O'Keefe, 2. J. Dostrovsky , Brain Res. 34, 171 (1971). [OpenUrl][14][CrossRef][15][PubMed][16][Web of Science][17] 2. [↵][18]1. T. Eliav et al ., Science 372, eabg4020 (2021). [OpenUrl][19][Abstract/FREE Full Text][20] 3. [↵][21]1. M. W. Jung, 2. S. I. Wiener, 3. B. L. McNaughton , J. Neurosci. 14, 7347 (1994). [OpenUrl][22][Abstract/FREE Full Text][23] 4. [↵][24]1. J. O'Keefe, 2. N. Burgess , Nature 381, 425 (1996). [OpenUrl][25][CrossRef][26][PubMed][27][Web of Science][28] 5. [↵][29]1. A. A. Fenton et al ., J. Neurosci. 28, 11250 (2008). [OpenUrl][30][Abstract/FREE Full Text][31] 6. [↵][32]1. P. D. Rich, 2. H.-P. Liaw, 3. A. K. Lee , Science 345, 814 (2014). [OpenUrl][33][Abstract/FREE Full Text][34] 7. [↵][35]1. K. B. Kjelstrup et al ., Science 321, 140 (2008). [OpenUrl][36][Abstract/FREE Full Text][37] 8. [↵][38]1. B. C. Harland et al ., bioRxiv 10.1101/2021.02.15.431172 (2021). 9. [↵][39]1. T. Hafting, 2. M. Fyhn, 3. S. Molden, 4. M.-B. Moser, 5. E. I. Moser , Nature 436, 801 (2005). [OpenUrl][40][CrossRef][41][PubMed][42][Web of Science][43] 10. [↵][44]1. V. H. Brun et al ., Hippocampus 18, 1200 (2008). [OpenUrl][45][CrossRef][46][PubMed][47][Web of Science][48] 11. [↵][49]1. C. J. MacDonald, 2. K. Q. Lepage, 3. U. T. Eden, 4. H. Eichenbaum , Neuron 71, 737 (2011). [OpenUrl][50][CrossRef][51][PubMed][52][Web of Science][53] [1]: #ref-1 [2]: #ref-2 [3]: #ref-3 [4]: #ref-4 [5]: #ref-6 [6]: #ref-5 [7]: #ref-7 [8]: pending:yes [9]: #ref-8 [10]: #ref-9 [11]: #ref-10 [12]: #ref-11 [13]: #xref-ref-1-1 "View reference 1 in text" [14]: {openurl}?query=rft.jtitle%253DBrain%2Bresearch%26rft.stitle%253DBrain%2BRes%26rft.aulast%253DO%2527Keefe%26rft.auinit1%253DJ.%26rft.volume%253D34%26rft.issue%253D1%26rft.spage%253D171%26rft.epage%253D175%26rft.atitle%253DThe%2Bhippocampus%2Bas%2Ba%2Bspatial%2Bmap.%2BPreliminary%2Bevidence%2Bfrom%2Bunit%2Bactivity%2Bin%2Bthe%2Bfreely-moving%2Brat.%26rft_id%253Dinfo%253Adoi%252F10.1016%252F0006-8993%252871%252990358-1%26rft_id%253Dinfo%253Apmid%252F5124915%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [15]: /lookup/external-ref?access_num=10.1016/0006-8993(71)90358-1&link_type=DOI [16]: /lookup/external-ref?access_num=5124915&link_type=MED&atom=%2Fsci%2F372%2F6545%2F913.atom [17]: /lookup/external-ref?access_num=A1971K844700012&link_type=ISI [18]: #xref-ref-2-1 "View reference 2 in text" [19]: {openurl}?query=rft.jtitle%253DScience%26rft.stitle%253DScience%26rft.aulast%253DEliav%26rft.auinit1%253DT.%26rft.volume%253D372%26rft.issue%253D6545%26rft.spage%253Deabg4020%26rft.epage%253Deabg4020%26rft.atitle%253DMultiscale%2Brepresentation%2Bof%2Bvery%2Blarge%2Benvironments%2Bin%2Bthe%2Bhippocampus%2Bof%2Bflying%2Bbats%26rft_id%253Dinfo%253Adoi%252F10.1126%252Fscience.abg4020%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [20]: /lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6Mzoic2NpIjtzOjU6InJlc2lkIjtzOjE3OiIzNzIvNjU0NS9lYWJnNDAyMCI7czo0OiJhdG9tIjtzOjIyOiIvc2NpLzM3Mi82NTQ1LzkxMy5hdG9tIjt9czo4OiJmcmFnbWVudCI7czowOiIiO30= [21]: #xref-ref-3-1 "View reference 3 in text" [22]: {openurl}?query=rft.jtitle%253DJournal%2Bof%2BNeuroscience%26rft.stitle%253DJ.%2BNeurosci.%26rft.aulast%253DJung%26rft.auinit1%253DM.%26rft.volume%253D14%26rft.issue%253D12%26rft.spage%253D7347%26rft.epage%253D7356%26rft.atitle%253DComparison%2Bof%2Bspatial%2Bfiring%2Bcharacteristics%2Bof%2Bunits%2Bin%2Bdorsal%2Band%2Bventral%2Bhippocampus%2Bof%2Bthe%2Brat%26rft_id%253Dinfo%253Adoi%252F10.1523%252FJNEUROSCI.14-12-07347.1994%26rft_id%253Dinfo%253Apmid%252F7996180%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [23]: /lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6Njoiam5ldXJvIjtzOjU6InJlc2lkIjtzOjEwOiIxNC8xMi83MzQ3IjtzOjQ6ImF0b20iO3M6MjI6Ii9zY2kvMzcyLzY1NDUvOTEzLmF0b20iO31zOjg6ImZyYWdtZW50IjtzOjA6IiI7fQ== [24]: #xref-ref-4-1 "View reference 4 in text" [25]: {openurl}?query=rft.jtitle%253DNature%26rft.stitle%253DNature%26rft.aulast%253DO%2527Keefe%26rft.auinit1%253DJ.%26rft.volume%253D381%26rft.issue%253D6581%26rft.spage%253D425%26rft.epage%253D428%26rft.atitle%253DGeometric%2Bdeterminants%2Bof%2Bthe%2Bplace%2Bfields%2Bof%2Bhippocampal%2Bneurons.%26rft_id%253Dinfo%253Adoi%252F10.1038%252F381425a0%26rft_id%253Dinfo%253Apmid%252F8632799%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [26]: /lookup/external-ref?access_num=10.1038/381425a0&link_type=DOI [27]: /lookup/external-ref?access_num=8632799&link_type=MED&atom=%2Fsci%2F372%2F6545%2F913.atom [28]: /lookup/external-ref?access_num=A1996UN47900052&link_type=ISI [29]: #xref-ref-5-1 "View reference 5 in text" [30]: {openurl}?query=rft.jtitle%253DJournal%2Bof%2BNeuroscience%26rft.stitle%253DJ.%2BNeurosci.%26rft.aulast%253DFenton%26rft.auinit1%253DA.%2BA.%26rft.volume%253D28%26rft.issue%253D44%26rft.spage%253D11250%26rft.epage%253D11262%26rft.atitle%253DUnmasking%2Bthe%2BCA1%2BEnsemble%2BPlace%2BCode%2Bby%2BExposures%2Bto%2BSmall%2Band%2BLarge%2BEnvironments%253A%2BMore%2BPlace%2BCells%2Band%2BMultiple%252C%2BIrregularly%2BArranged%252C%2Band%2BExpanded%2BPlace%2BFields%2Bin%2Bthe%2BLarger%2BSpace%26rft_id%253Dinfo%253Adoi%252F10.1523%252FJNEUROSCI.2862-08.2008%26rft_id%253Dinfo%253Apmid%252F18971467%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [31]: /lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6Njoiam5ldXJvIjtzOjU6InJlc2lkIjtzOjExOiIyOC80NC8xMTI1MCI7czo0OiJhdG9tIjtzOjIyOiIvc2NpLzM3Mi82NTQ1LzkxMy5hdG9tIjt9czo4OiJmcmFnbWVudCI7czowOiIiO30= [32]: #xref-ref-6-1 "View reference 6 in text" [33]: {openurl}?query=rft.jtitle%253DScience%26rft.stitle%253DScience%26rft.aulast%253DRich%26rft.auinit1%253DP.%2BD.%26rft.volume%253D345%26rft.issue%253D6198%26rft.spage%253D814%26rft.epage%253D817%26rft.atitle%253DLarge%2Benvironments%2Breveal%2Bthe%2Bstatistical%2Bstructure%2Bgoverning%2Bhippocampal%2Brepresentations%26rft_id%253Dinfo%253Adoi%252F10.1126%252Fscience.1255635%26rft_id%253Dinfo%253Apmid%252F25124440%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [34]: /lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6Mzoic2NpIjtzOjU6InJlc2lkIjtzOjEyOiIzNDUvNjE5OC84MTQiO3M6NDoiYXRvbSI7czoyMjoiL3NjaS8zNzIvNjU0NS85MTMuYXRvbSI7fXM6ODoiZnJhZ21lbnQiO3M6MDoiIjt9 [35]: #xref-ref-7-1 "View reference 7 in text" [36]: {openurl}?query=rft.jtitle%253DScience%26rft.stitle%253DScience%26rft.aulast%253DKjelstrup%26rft.auinit1%253DK.%2BB.%26rft.volume%253D321%26rft.issue%253D5885%26rft.spage%253D140%26rft.epage%253D143%26rft.atitle%253DFinite%2BScale%2Bof%2BSpatial%2BRepresentation%2Bin%2Bthe%2BHippocampus%26rft_id%253Dinfo%253Adoi%252F10.1126%252Fscience.1157086%26rft_id%253Dinfo%253Apmid%252F18599792%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [37]: /lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6Mzoic2NpIjtzOjU6InJlc2lkIjtzOjEyOiIzMjEvNTg4NS8xNDAiO3M6NDoiYXRvbSI7czoyMjoiL3NjaS8zNzIvNjU0NS85MTMuYXRvbSI7fXM6ODoiZnJhZ21lbnQiO3M6MDoiIjt9 [38]: #xref-ref-8-1 "View reference 8 in text" [39]: #xref-ref-9-1 "View reference 9 in text" [40]: {openurl}?query=rft.jtitle%253DNature%26rft.stitle%253DNature%26rft.aulast%253DHafting%26rft.auinit1%253DT.%26rft.volume%253D436%26rft.issue%253D7052%26rft.spage%253D801%26rft.epage%253D806%26rft.atitle%253DMicrostructure%2Bof%2Ba%2Bspatial%2Bmap%2Bin%2Bthe%2Bentorhinal%2Bcortex.%26rft_id%253Dinfo%253Adoi%252F10.1038%252Fnature03721%26rft_id%253Dinfo%253Apmid%252F15965463%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [41]: /lookup/external-ref?access_num=10.1038/nature03721&link_type=DOI [42]: /lookup/external-ref?access_num=15965463&link_type=MED&atom=%2Fsci%2F372%2F6545%2F913.atom [43]: /lookup/external-ref?access_num=000231116500035&link_type=ISI [44]: #xref-ref-10-1 "View reference 10 in text" [45]: {openurl}?query=rft.jtitle%253DHippocampus%26rft.stitle%253DHippocampus%26rft.aulast%253DBrun%26rft.auinit1%253DV.%2BH.%26rft.volume%253D18%26rft.issue%253D12%26rft.spage%253D1200%26rft.epage%253D1212%26rft.atitle%253DProgressive%2Bincrease%2Bin%2Bgrid%2Bscale%2Bfrom%2Bdorsal%2Bto%2Bventral%2Bmedial%2Bentorhinal%2Bcortex.%26rft_id%253Dinfo%253Adoi%252F10.1002%252Fhipo.20504%26rft_id%253Dinfo%253Apmid%252F19021257%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [46]: /lookup/external-ref?access_num=10.1002/hipo.20504&link_type=DOI [47]: /lookup/external-ref?access_num=19021257&link_type=MED&atom=%2Fsci%2F372%2F6545%2F913.atom [48]: /lookup/external-ref?access_num=000261871800006&link_type=ISI [49]: #xref-ref-11-1 "View reference 11 in text" [50]: {openurl}?query=rft.jtitle%253DNeuron%26rft.volume%253D71%26rft.spage%253D737%26rft_id%253Dinfo%253Adoi%252F10.1016%252Fj.neuron.2011.07.012%26rft_id%253Dinfo%253Apmid%252F21867888%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [51]: /lookup/external-ref?access_num=10.1016/j.neuron.2011.07.012&link_type=DOI [52]: /lookup/external-ref?access_num=21867888&link_type=MED&atom=%2Fsci%2F372%2F6545%2F913.atom [53]: /lookup/external-ref?access_num=000294521600017&link_type=ISI

Since the starting of evolution of Artificial Intelligence (AI) in the world, people are curious about various things related to technology. One such major question is, 'Will artificial intelligence have a conscience?' Even though when scientists provide diverse answers from their perspective, it is still a lingering thought to many. Talking constantly about AI conscience may sound out of the box now, but people should keep in mind that AI is improving every day. Soon, AI is expected to care for the elderly, teach children and perform many tasks that require moral human judgement. Henceforth, it has the necessity to differentiate between what is right and wrong.

Researchers at the University of Zurich's Brain Research Institute have recently developed a technique to automatically detect neurons of different types in a variety of brain regions at different developmental stages. They presented this deep learning-based tool, called DeNeRD, in a paper published in Nature Scientific Reports. Mapping the structure of the mammalian brain at the cellular level is an important, yet demanding task, which typically involves capturing specific anatomical features and analyzing them. In the past, researchers were able to gather several interesting observations and insights about the mammalian brain's structure using classical histological and stereological techniques. Although these methods have proved to be very useful for studying the anatomy of the brain, carrying out a truly brain-wide analysis typically requires a different approach.

Scientists are about to get a serious assist in their quest to simulate brains. HPE has deployed Blue Brain 5, a supercomputer dedicated to simulations and reconstructions of mammalian brains as part of the École Polytechnique Fédérale de Lausanne's Blue Brain Project. The system is based on HPE's existing SGI 8600 (above) and packs a hefty 372 compute nodes between its Xeon Gold, Xeon Phi and Tesla V100 processors, not to mention a whopping 94TB of memory. More importantly, it's flexible -- Blue Brain 5 has four configurations to prioritize different computing tasks, and it can host subsystems geared toward relevant tasks (including deep learning and visualization) while operating as a cohesive whole. This kind of power is necessary, even if simulating a complete brain is still a long ways off.

Call it an a-MAZE-ing development: A U.K.-based team of researchers has developed an artificial intelligence program that can learn to take shortcuts through a labyrinth to reach its goal. In the process, the program developed structures akin to those in the human brain. The emergence of these computational "grid cells," described in the journal Nature, could help scientists design better navigational software for future robots and even offer a new window through which to probe the mysteries of the mammalian brain. In recent years, AI researchers have developed and fine-tuned deep-learning networks -- layered programs that can come up with novel solutions to achieve their assigned goal. For example, a deep-learning network can be told which face to identify in a series of different photos, and through several rounds of training, can tune its algorithms until it spots the right face virtually every time.

It has been put to good use by comic book superheroes and by alien predators hell-bent on wiping out mankind, but soon humans could also be able to see infrared light. Scientists have used brain implants to give rats a'sixth-sense' that enables them to detect and react to the normally invisible light source. The research proves it is possible for the adult brain to adapt to new forms of input and opens up the possibility of enabling humans to gain an array of superhuman senses. Scientists have connected infrared sensors to the brains of rats using electrical implants to allow the rodents to detect the normally invisible light. They found the rats were able to spot infrared light and react to it by pressing a button beneath an infrared source to get food. Researchers say it may be possible to attach sensors for other forms of light such as ultraviolet, microwaves and even x-rays using brain implants.