FR2920582A1 - Human language comprehension device for robot in e.g. medical field, has supervision and control system unit managing and controlling functioning of device in group of anterior information units and electrical, light and chemical energies - Google Patents

Abstract

The device (1) has a supervision and control system unit (2) for managing and controlling functioning of the device in a group of anterior information units (4) and managing and controlling electrical, light and chemical energies. The supervision and control system unit includes input sources placed on information units (10) of a group of immediate units contained in a comprehension unit (3). The group of anterior information units includes a variable number of the information units.

Description

The understanding of human language, whether spoken, written, based on images, movements or gestures, involves information of perceptions and human reactions previously recorded and organized in memory. Currently, devices or devices that may include some form of human communication are limited in scope to a specific area and are often of constraining use. Sometimes they are limited to reacting to a limited vocabulary previously defined or some expressions fixed in advance. Internet search engines recognize a few key words. Dictation or speech recognition software does not know the meaning of words, phrases, sentences, is not good for noise, and often only recognizes the voices of interlocutors previously recorded. Automatic language processing is so far limited to very specialized fields, with a limited vocabulary. Sometimes robots can learn to recognize people, places, objects, read handwritten texts, or react to particular situations. But their learning is slow and their possibilities remain marginal or very special. Neural networks know how to recognize shapes, faces, people, read certain handwritten characters, however their development can be laborious and more art than science to define their structure, their neuronal weights, the samples test. The present device proposes to exploit information preserved after having been previously encountered, corresponding to human visual or sound perceptions, or to visual or sound actions of the device, to understand the human language in its generality and to prepare reactions or adapted answers.

Operating principle of the present device. The device is activated upon activation of the Supervision and Control system unit. This unit provides the operating energy of the device, manages and controls the device as a whole, manages and controls the energy that can be electrical, light, chemical. The apparatus on which the device is mounted comprises a unit that supplies input sources with all the immediate information units of the device's understanding unit. Input sources are information that corresponds to human visual or sound perceptions.

The concentration energy of the Supervision and Control system unit chronologically breaks down the input sources into elementary information signals, on the information portions of the Time Support units of the information units of the set of information elements. Immediate Information units. The energy of concentration connects together the elementary information signals which are revealed by their characteristics, in order to form spatiotemporal coded sets, either in space on a unit of temporal support of information elements, or in space and time on several sequential units of temporal support of information elements. A visual framework, corresponding to what a human being can see in a unit of time, can be broken down into several codified sets. A sound frame, corresponding to what a human being can hear in a unit of time, can be broken down into several codified sets. A visual action frame, corresponding to visual actions that the device can perform in a unit of time, can be broken down into several coded sets. A sound action frame, corresponding to sound actions that the device can perform in a unit of time, can be broken down into several coded sets. A sequence of information, corresponding to a chronological sequence, where visual frames, sound frames, visual action frames, sound action frames can be mixed, can be broken down into several units of temporal support of elements. of information in several codified spatio-temporal sets. Codified sets can themselves be combined to form new coded sets. A subset of elementary information signals of a coded set may form a new coded set. A syllable, a sound, a voice, a word, a firm, a criterion, a word, a diagram, a movement, a look, an object, an approach, a person, a concept, an expression, a sentence, a rule , a route, an apartment, a village, can for example be defined as codified sets. The same word, the same object, the same person, the same concept, can for example be defined by multiple codified sets. Thus, a word can be defined by codified sets visual, sound, visual actions, sound actions, its display, its listening, its writing, its pronunciation, its inflections and variations, like the kind , the number, the conjugation, the elision, the connections. The codified sets are associated with sequences of other codified sets, which can determine for example their meanings, their use cases, their grammatical uses. Some of these codified sets of sequences reflect the circumstances in which definitions or uses have taken place, and the coded sets of these adjacent sequences may participate in the search for the coded sets to which they have been associated during the process of their comprehension. or their identification. Some coded sets correspond to markers of the language such as grammatical markers, logical operators, mathematical operators, grammar rules, logic rules, mathematical functions.

The coded sets found on the Information units of all Immediate Information units are highlighted by an electrical, luminous, chemical process, which makes them visible simultaneously by all Contextual Information units. by all previous units of Information. The set of Contextual Information units contains the codified sets relating to comprehension processes recently carried out, which may concern current events, recent events or up-to-date information such as place, date. The set of Prior Information units contains coded sets of information that have been preserved after having been previously placed, during a prior understanding process, on these units when the latter were first part of the set of Immediate Information Units and then is part of all Contextual Information Units. The set of Prior Information units comprises, on the information parts of its Time Support units, information initialized to the construction of the device, corresponding to human visual perceptions, to human sound perceptions, to visual actions. of the device, to sound actions of the device, and to the links of their understanding, which may include, at varying levels, one or more dictionaries, one or more grammars, logical rules, mathematical operations, mathematical functions, information specialized. The identification phase of the coded sets on the Immediate Information units is started as the coded sets are formed. The Supervision and Control system unit can also launch more specific searches from certain coded sets, to collect information previously encountered. This search is carried out in parallel on the information units of all the Contextual Information units and on the information units of all the previous information units, under control of the control systems specific to these units. Information. The same search can involve a variable number of codified sets. These searches can take into account or may not take into account variable criteria such as functions (identity, homothety, symmetry, stretching of shapes, ...), Boolean or logical or mathematical operations (yes, no, equal, no equal, and, or inclusive, or exclusive,

.), properties of codified sets (colors, shapes, movements, steps, orientations, voices, voice stamps, voice intensities, accents, lyrics, pronunciation mistakes, ...). The searches are flexible because, on the one hand, a requested criterion can go from the value corresponds perfectly to the value does not correspond at all, with intermediate intermediate values of partial correspondences and, on the other hand, the requested criteria can go of the number all the criteria correspond to the number no criterion corresponds, with a variable intermediate number of corresponding criteria. The information unit control systems of the set of Contextual Information Units and all the prior information units can select the information element time support units containing coded sets previously encountered, meeting the requested search criteria, and the information element temporal support units containing the coded sets associated with them during a prior process of understanding, highlighting the coded sets involved. Connection links are established between the information element temporal support units, the set of Contextual Information units and all the prior information units, containing the coded sets responding to the searches and the associated coded sets, and the units of the Immediate Information set where the coded sets to be identified or understood are located, under the control of the control systems specific to the information units that contain them. These connection links, which can be electrical, luminous, chemical, vary in intensity, ranging from very strong to very low, depending on the results of research that go from very positive to very little positive, and they differentiate the results of research in separate unit groups ... DTD: The Supervision and Control system unit examines the search results and prioritizes the information element temporal support unit groups • with connection links of higher intensity; • having coded sets that are found in multiple occurrences; Having an acceptable combination of coded sets corresponding to markers of the language. The groups of temporal support units of information elements chosen after examination of the results of the searches are placed on the mechanism of representation and internal projection of the information of Perceptions and Actions, under control of the unit 10 system of Supervision and of Control, where the parts of information corresponding to visual perceptions, to sound perceptions, to visual actions of the device, to sound actions of the device, and to the links of Understanding, will be placed respectively on the Representation internal information corresponding to Visual Perceptions, Sound Perceptions, Visual Actions of the device, Sound Actions of the device, and Internal Representation of the Understanding links. The mechanism of Internal Representation and Projection of Perceptions and Actions information, is an internal reproduction mechanism sensitive to reactions that can be electrical, luminous, chemical. It takes up the spatio-temporal information of the codified sets chosen and sequences of associated coded sets, to project them into an internal space, to restore their orientation, their positions, their relief, to animate them; to reproduce them in terms of images and sounds; to create new associations. It represents the device itself as being an active observer of the connections, built in its internal universe, of information such as 25 objects, persons, experiences, situations, notions. This mechanism exploits the Internal Representation of the Comprehension Links corresponding to the Connections of Understanding that have been codified in a previous process of understanding. From this mechanism, the Supervision and Control system unit extracts or creates new coded sets, places them on units of the set of Immediate Information, makes them visible to the units of all the units of the set. Contextual Information and all previous information units, and launches new research on the units of these sets. For example, searches can be started • in the set of Contextual Information units, to find contextual values that correspond to extracted coded sets, • in the set of Prior Information units, to find codified sets models associated with the operation of one or more extracted language markers, which specify the rules for replacing model values by current values, • in the set of Prior Information units, to find complementary coded sets, starting from the creation of new codified sets, which specify variable elements, aggregate sequences, create new internal associations. The information element temporal support units, containing the coded sets to be studied on the information units of the set of immediate information units, are also placed on the mechanism of representation and internal projection of the information elements. Perceptions and Actions.

Under the control of the Supervision and Control system unit, the Internal Projection of the Understanding information of the mechanism of Representation and Internal projection of the information of Perceptions and Actions, animates internal visual sequences, internally displays the information to identify or understand, internally display the information identified or understood, from the information corresponding to Visual Perception or Visual Actions of the device; • animates internal sound sequences, internally pronounces the information to be identified or understood, internally pronounces the information identified or understood, from the information corresponding to Sound Perception or Sound Actions of the device; • links together, on the Internal Representation of Comprehension Links, codified sets relating to identification or comprehension; • Places, in coded form, the Understanding links thus created on the link information part of the Understanding of the units involved of all the Immediate Information units. The Supervision and Control system unit manages the comprehension process by comparing, on the Internal projection of the Understanding information of the mechanism of Representation and Internal projection of the information of Perceptions and Actions, the links that have just been created on the Internal Representation of the Comprehension Links, from the coded sets of the current information sequence to be identified or from the current sentence proposal to be understood; • the links that have just been created on the Internal Representation of the Comprehension Links, from the coded sets chosen to identify the current sequence of information or to understand the current sentence proposition; • The comprehension links from the information part of the links of the Understanding of all Contextual Information units, relative to the last sentence proposals, to the last sentences, to the last information sequences, after their reactivation. If, during the control of the comprehension process, the Supervision and Control system unit encounters an anomaly of comprehension, or likelihood, or coherence, it may choose other groups of units of Temporary Support from information items satisfying the searches, and it re-runs the process described above, from the selection of these groups of units of Temporary Support. If, during the control of the comprehension process, the Supervision and Control system unit encounters an understanding, or likelihood, or coherence discrepancy, and there are no more Support unit groups Temporary of information elements satisfying research, it takes again the process of understanding, since the constitution of the initial codified sets, constituting different coded sets, or taking other search criteria. If the Supervision and Control system unit still encounters an anomaly of comprehension, or likelihood, or coherence, it positions coded sets of action requests corresponding to the type of anomaly and links these coded sets to the links of Comprehension. When the Supervision and Control system unit detects a request for action or a request for a response, based on the Comprehension links and associated coded sets, it initiates searches in all of the units. Contextual Information or in all previous units of information; • it places the results on the mechanism of Representation and Internal projection of Perceptions and Actions information; • it selects the coded sets relating to the meaning of the answer or relating to the communication of the action, on the Internal Representation of the Comprehension links and on the Internal Representation of the information corresponding to Visual Actions of the device or to Sound Actions the device; It transforms the preceding information into information codified on the part of information corresponding to visual actions of the device and to sound actions of the device of the Temporary Information Element Support Units of the units of Information involved; • it updates the corresponding Understanding links; • it makes, either immediately or on the basis of events, the output actions information, thus prepared on all the Immediate Information units, available to the unit of the output actions of the device on which is mounted the device When the Supervision and Control system unit considers the understanding of a sentence proposal or a sequence of information to be satisfactory, • it dissociates the information units from all the Immediate Information units. containing the coded sets and their associated Understanding links of all Immediate Information units to form part of all Contextual Information units; • and allocates new information units to all Immediate Information units. At a variable interval, under the control of the Supervision and Control system unit, the information units of the set of Contextual Information units are dissociated from the comprehension unit to be part of all the units. Previous information. During this dissociation, the concerned information units retain a variable part of their information, which may include the new codified sets, the new codified sets associated with them, and the coded sets linked by the Comprehension links.

The accompanying drawings show operating diagrams of the present invention. FIG. 1 shows the unit of the input sources (1A), the unit of the output actions (1B). These units (1A) and (1B) communicate with the present device (1) and are part of the device that has this device (1). • Figure 2 shows that the present device (1) comprises the Supervision and Control system unit (2), the Comprehension unit (3), the set of Prior Information units (4). The set of Prior Information units (4) comprises a variable number of the information unit (10). • Figure 3 shows that the comprehension unit (3) comprises all the immediate information units (5), the set of Contextual Information units (6), the mechanism of Representation and Internal Projection of Perceptions and Actions information (7). The set of Immediate Information units (5) comprises a variable number of the information unit (10).

The set of Contextual Information units (6) comprises a variable number of the information unit (10). Figure 4 shows that the information unit (10) comprises the control system (20), a variable number of the information element temporal support unit (30). Figure 5 shows that the information element temporal support unit (30) comprises: the information portion corresponding to visual perceptions (31); the information portion corresponding to sound perceptions (32); ), the information part corresponding to the links of the understanding (34), - the information part corresponding to visual actions of the device (35), - the part of information corresponding to sound actions of the device (36) . FIG. 6 shows that the information portion corresponding to visual perceptions (31) of information element temporal support units (30) comprises elementary information signals (40), coded sets (41), ) spatio-temporal constituting elementary information signals linked in space and on the same unit of temporal support of information elements (30), coded sets (41) spatio-temporal consisting of elementary signals d information relating to space and to a plurality of information element (30) temporal support units. • Figure 7 shows the set of Immediate Information units (5), all Contextual Information units (6), all Previous Information units (4). The set of Immediate Information units (5) contains a coded set (41). The set of Prior Information units (4) contains coded sets (41) which may be - occurrences of the coded set (41) found on the set of Immediate Information units (5); coded sets (41) associated by language markers; coded sets (41) of an associated information sequence; codified sets (41) others. • Figure 8 shows that the mechanism of Internal Representation and Projection of Perceptions and Actions information (7) includes the Internal Representation of Information corresponding to Visual Perceptions (71) to Sound Perceptions (72), to Visual Actions. from the device (75), to sound actions of the device (76); - Internal projection of the Understanding information (70); - The Internal Representation of the Understanding Links (74).

Description of an implementation The device that supports the device (1) is turned on. Activation of the device (1) activates its units. A person comes to the device. Input sources (1A) corresponding to visual and sound perceptions are recorded in the apparatus. The supervisory and control system unit (2) of the device (1) will take into account the input sources (1A) and place them on the information units (10) of all the information units. Immediate (5) contained in the comprehension unit (3). The concentration energy of the supervisory and control system unit (2) chronologically decomposes the input sources (1A) into elementary information signals (40) on time element units of information elements (30) Information units (10) of all Immediate Information units (5). Concentration energy connects together elementary information signals that are found to be close by their characteristics, to form spatiotemporally coded sets (41), or in space on a unit of temporal support of information elements. (30), ie in space and time on several sequential units of information element temporal support (30). Thus, codified sets (41) of the person's gait, body, clothes, head, hair, eyes, smile, ... can be formed on the part corresponding to visual perceptions (31) of the information element temporal support units (30); simultaneously, codified sets (41) of the person's voice tone, the pitch of his voice, his voice modulations, his accent, the sounds of his words, ... can be formed on the part corresponding to sound perceptions (32) of the information element temporal support units (30). As these coded sets (41) are formed, they are made visible by the set of Contextual Information units (6) and by the set of Prior Information units (4). The supervisory and control system unit (2) initiates correspondence searches which run in parallel on the information units (10) of these sets, each of these information units (10) being under the control of his own control system (20). An already registered person may come with some aspects different from those registered: she may have changed clothes or hairstyle, injured her face or leg, wear sunglasses, have a hoarse voice, etc. Also, each search criterion is flexible, ie values that only partially match can be accepted; and the set of search criteria combine flexibly, that is, it may be acceptable that some of the desired values do not match.

Control systems (20) managing the information units (10) establish connection links between the information element time support unit groups (30) of the information units (10) of the information units (10) of the information units (10) of the information units (10) of the information units set of Contextual Information Units (6) and all the Previous Information units (4), containing coded sets (41) corresponding to the searches and coded sets (41) associated with them, and the units information element (30) of the Information units (10) of the Immediate Information set (5) where the coded sets (41) of the person to be identified are located. These bright and chemical electrical connection links vary in intensity according to the results of the research. The control systems (20) differentiate the search results into groups of distinct units and highlight the coded sets (41) involved. The supervisory and control system unit (2) examines the search results and first focuses on the information element temporal support unit groups (30) whose connection links are of greater intensity: those that best meet the search criteria; then those for which several criteria are satisfied in multiple occurrences. The supervision and control system unit (2) places the groups of temporal support units of selected information elements (30) after examination of the search results containing the coded sets (41) searched for and those which are associated with the mechanism of representation and internal projection of the information of Perceptions and Actions (7), where • the parts of information corresponding to visual perceptions (31), to sound perceptions (32), to visual actions of the device (35), to sound actions of the device (36), to the links of the Comprehension (34), • will be placed respectively on the Internal Representation of the information corresponding to Visual Perceptions (71), to Perceptive Sounds (72) , to Visual Actions of the device (75), to Sound Actions of the device (76), on the Internal Representation of the Comprehension Links (74).

The supervisory and control system unit (2) also places the information element temporal support units (30) containing the coded sets to identify information units (10) of the set of units. Immediate Information (5) on the Representation and Internal Projection Mechanism of Perceptions and Actions Information (7).

The Mechanism of Representation and Internal Projection of the Perceptions and Actions Information (7), takes up the spatio-temporal information of the selected coded sets (41) and the coded sets of sets (41) associated to project them, on the Projection part internal understanding information (70), internal space sensitive to light and chemical electrical reactions. He animates sequences of internal information visual and sound. It links between them, on the Internal Representation of the Understanding Links (74), the coded sets (41) of the current information sequence to be identified and the coded sets (41) found and likely to identify the person, and place these links in coded form on the link information portion of the Understanding (34) of the Information Element Time Support Units (30) of all Immediate Information units (5) involved in the identification of the person. The supervisory and control system unit (2) controls the identification process by comparing, on the internal projection of the understanding information (70) of the representation and internal projection mechanism of the perception and action information (7), the Understanding links (74) created from the coded sets (41) of the current information sequence to be identified, and the Comprehension links (74) created from the coded sets (41) chosen to identify the sequence of current information. If the comparison gives a positive result, the Supervision and Control system unit (2) • considers that it has found a person already met previously, • searches in the Temporary Support units of information elements (30). ) associated with the codified sets (41) satisfying the searches if there are codified sets (41) which make it possible to identify the person, such as his written name, his spoken name, his first name, his place of residence. If the comparison gives a null or not very positive result to identify the person, such as for example only the same clothing or hairstyle, the unit Supervision and Control system (2) considers that it is a new person . After completing the person identification process, the Supervision and Control system unit (2) dissociates the information units (10) from all the Immediate Information units (5) containing the sets. codified (41) and the associated Understanding (34) links of all the Immediate Information units (5) to form part of the set of Contextual Information units (6); and allocates new information units (10) to all Immediate Information units (5). The person expresses himself, makes gestures and moves. The device (1) receives input sources (1A) that correspond to the visual perceptions of his words, his movements, his gestures, to the sound perceptions of his words, and also to the frames of visual and sound perceptions where the nobody. The concentration energy of the supervisory and control system unit (2) chronologically decomposes the input sources (1A) into elementary information signals (40) on time element units of information elements (30) Information units (10) of all Immediate Information units (5) contained in the Understanding Unit (3). The supervisory and control system unit (2) centralizes its energy on coded sets (41) which it constitutes from the elementary information signals (40) which correspond to the visual and sound perceptions of the words, movements and gestures of the person. White coded sets (41) can also be formed from the person's language breaks. As these coded sets (41) are constituted, they are made visible by the set of Contextual Information units (6) and by the set of Prior Information units (4) and the unit Supervision and Control system (2) initiates correspondence searches that run in parallel on the information units (10) of these sets, each unit being under the control of its own control system (20).

The search criteria are not limited to codified sets (41) sound corresponding to the words spoken by the person, but they also include codified sets (41) visual corresponding to movements of the mouth, facial expressions, gestures, ... of the person speaking. If the person has already been previously recorded by the device (1), visual or sound sequences of codified sets (41) can be found, corresponding to mimicry, expressions, turns of language that characterize it. The control systems (20) of the information units (10) establish connection links between the information element time support units (30) of the set of Immediate Information (5) where the coded sets (41) to be identified and the information element temporal support unit groups (30) of all the Contextual Information units (6) and all the prior information units ( 4) which are associated with the coded sets (41) satisfying the searches, highlighting the coded sets (41) which correspond to the searches and, among others, the coded sets (41) associated with their grammatical natures (name, adjective, verb, article, pronoun, adverb, preposition, conjunction, expression, idiocy, quote, ...). The supervisory and control system unit (2) examines the search results, first selecting the information element (30) time support unit groups that have the strongest connection links. intensity. The selected choice includes, among others, the codified set (41) corresponding to the ilplodekord sound perception, which includes, among its associated coded sets (41), the coded sets (41) corresponding to the sound perceptions .

The Supervision and Control system unit (2) places on the Internal Representation and Projection mechanism the Perception and Actions information (7) of the Understanding Unit (3). information (30) containing the coded sets (41) associated with the preceding choice; And the information element temporal support units (30) of the information units (10) of the immediate information set (5), containing the coded sets (41) to be identified and understood. Under the control of the Supervision and Control system unit (2), the Internal Projection of the Comprehension information (70) • internally display information, derived from the associations, corresponding to Visual Perceptions (71) and Visual Actions of the device (75): - grammatical information: expression, consisting of the following: pronoun, verb, determinant, noun; 15 - the text that has been understood: it's raining cats; a text equivalent to its meaning: it rains a lot; - a sequence of images of its meaning: images of heavy rain; • and at the same time hears internally the information, coming from the associations, corresponding to Sound Perceptions (72) and to Sound Actions of the device 20 (76): - the text which has been understood: it is raining cords; - a text equivalent to its meaning: it rains a lot. The supervisory and control system unit (2) • creates links to the coded sets (41) involved in the understanding phase of the expression, on the Internal Representation of the Comprehension Links (74); Places, in coded form, these links (74) created, on the link part (34) of the information element temporal support units (30) of the information units (10) of the set of the information units; 'Immediate Information (5); • finds, by comparing, on the Internal Projection of the Understanding information 30 (70), projections resulting from the links (74) created from the coded sets (41) of the expression to be understood and links (74) created to from the codified sets (41) chosen to understand the expression, that there is a correspondence; • considers that the expression is properly understood and forms a complete sentence proposition; • dissociates the information units (10) containing coded sets (41) involved in the understanding of the expression, from the set of Immediate Information units (5) to connect them to all the units of Contextual Information (6), allocating new information units (10) to all Immediate Information units (5). The understanding process continues and the Supervision and Control system unit (2) examines the following sound sequence: 3asDrtireplytar.

On the information element temporal support units (30) of the information units (10) of the set of immediate information units (5), different coded sets (41) are formed from the elementary signals. information (40) corresponding to the visual and sound perceptions. As they are formed, they are made visible by all the Contextual Information units (6) and by all the Previous Information units (4); and the supervisory and control system unit (2) initiates correspondence searches which run in parallel on the information units (10) of these sets, each unit being under control of its own control system (20). . The control systems (20) of the information units (10) establish connection links between the units of the set of Immediate Information (5) and the information element time support unit groups (5). 30) of the set of Contextual Information Units (6) and of all the Previous Information units (4) which are associated with the coded audio and visual sets (41) satisfying the searches, highlighting the codified sets (41) that are involved.

The supervision and control system unit (2) examines the results of the searches and comes to choose, among others, the coded sets (41) corresponding to the sound perceptions 3a, sDrtire, plytar, which have among their codified sets (41) those who correspond to the perceptions of soundness, ply, tan, and those which correspond to their grammatical natures.

The Supervision and Control system unit (2) places on the Internal Representation and Projection mechanism the Perception and Actions information (7) of the Understanding Unit (3). information (30) containing the selected coded sets (41) and the coded sets (41) associated therewith; And the information element temporal support units (30) of the information units (10) of the immediate information set (5), containing the coded sets (41) to be identified and understood. Under the control of the Supervision and Control system unit (2), the mechanism of Internal Representation and Projection of the Perceptions and Actions information (7) • internally displays the sequence of the codified sets (41) grammatical nature: pronoun , verb, expression, associated with the selected coded sets (41); • launches a search in all Contextual Information units (6) to substitute for the codified set (41) pronoun: I, associated with the coded set (41): 3a, its contextual value: the identified person who speaks; • displays the coded set (41): will exit, associated with the coded set (41) selected: s3rtire; it is a temporal marker of the future, associated with the codified set (41): sDrtir, verb: to go out; • displays the coded set (41): later, associated with the selected coded set: plytar, to which the coded sets (41): ply, tan> are also associated; it marks a vague future. Then, the Internal Projection of Comprehension Information (70) • internally displays information, resulting from the associations, corresponding to Visual Perceptions (71) and Visual Actions of the device (75): - the text that has been understood: I will go out later; - a text equivalent to its meaning: at another future time, I, the person who speaks, I will go outside; - a sequence of images of its meaning: the hour that passes, then the person who goes outside; • At the same time, the information coming from the associations, corresponding to Sound Perceptions (72) and Sound Actions from the device (76), is heard internally: - the text that has been understood: I will come out later; - a text equivalent to its meaning: at another future time, I, the person who speaks, I will go outside; The Supervision and Control system unit (2) • creates links to the coded sets (41) involved in the understanding phase of the expression, on the Internal Representation of the Comprehension Links (74); • places in coded form these links (74) created on the links part (34) of the information elements temporal support units (30) of the information units (10) of the set of immediate information units (5); • observes, by comparing, on the Internal projection of the Understanding information (70), the links (74) of the coded sets (41) chosen at the links (74) of the coded sets (41) initially formed, that the choice is acceptable; • reactivates, on the Internal Projection of Comprehension Information (70), the links recently placed on all Contextual Information units (6); • finds, by comparing the links (74) of the coded sets (41) chosen links (74) reactivated from the previous expression on all Contextual Information units (6), that the choice is acceptable; • note that the sequence of words chosen forms an acceptable sentence proposal; • considers this sentence proposal properly understood; • dissociates the information units (10) containing coded sets (41) involved in the understanding of the expression, from the set of Immediate Information units (5) to connect them to all the units of Contextual Information (6), and allocates new Information units (10) to all Immediate Information units (5). The Supervision and Control system unit (2) continues to examine the following sound sequence: prevjËmafij. As previously described, there is formation of codified sets (41), and searches.

The Supervision and Control system unit (2) examines the search results and comes to choose, among others, the coded sets (41) corresponding to the sound perceptions prevj, ma, fij. The Supervisory and Control System Unit (2) places on the Internal Representation and Projection mechanism the Perception and Actions information (7) of the Understanding Unit (3) • The Elements of Temporary Support Elements information (30) containing the selected coded sets (41) and the coded sets (41) associated with them, such as those corresponding to their grammatical natures; And the information element temporal support units (30) of the information units (10) of the immediate information set (5), containing the coded sets (41) to be identified and understood. Under the control of the Supervision and Control system unit (2), the mechanism of Internal Representation and Projection of Perceptions and Actions information (7) • internally displays the sequence of coded sets (41) grammatical nature: verb , determinant, noun, associated with the selected coded sets (41); • starts a search in all Contextual Information units (6) to replace the codified (41) possessive adjective set: ma, associated with the coded set (41): ma, its contextual value: from identified person who speaks; • searches all of the Contextual Information units (6) and all the Previous Information units (4) to find the coded set (41) corresponding to: daughter of the speaker; • displays the coded set (41): warns, associated with the coded set (41) chosen: prevjÈ; it is an imperative, associated with the codified set (41): prevanir, verb prevent, marker of a request: to warn the girl of the person who speaks.

The supervision and control system unit (2) seeks to identify the girl of the person speaking, by examining the results of the research: • if the person has already been met before, that she has a daughter, and that his daughter is known, some codified sets (41) associated with him can identify, locate, join; • if the person is known to have no children, a request for explanation is prepared on the Internal Representation of Understanding Links (74) and on the Visual Actions of the device (75) or on the Sound Actions of the device ( 76); this request is sent codified on the visual action parts of the device (35) or sound actions of the device (36) of the information element temporal support units (30) of the units (10) of the set of units of Immediate Information (5) concerned; then received by the unit (1B) which communicates the output actions; the process of understanding is put on hold until the person intervenes to give explanations or clarifications. • if the person is new or if the person is known with several known children, - a request for additional information about the girl the person is talking about is prepared as described above; and the comprehension process is also put on hold until the person intervenes to specify the coordinates of his daughter. The Internal Projection of Understanding Information (70) • internally displays information from associations, corresponding to Visual Perceptions (71) and Visual Actions of the device (75): the text that has been understood: warn my daughter ; - a text equivalent to its meaning: sends a message to Anne; a sequence of images of its meaning: a representation of the device that calls Anne; • and at the same time hears internally the information coming from the associations, corresponding to Sound Perceptions (72) and Sound Actions of the device (76): - the text that has been understood: warn my daughter; - a text equivalent to its meaning: sends a message to Anne; The Supervision and Control system unit (2) • creates links to the coded sets (41) involved in the understanding phase of the expression, on the Internal Representation of the Comprehension Links (74); • places in coded form these links (74) created on the links part (34) of the information elements temporal support units (30) of the information units (10) of the set of immediate information units (5); • finds, by comparing the links (74) of the coded sets (41) chosen at the links (74) of the coded sets (41) initially formed, that the choice is acceptable; • reactivates, on the Internal Projection of Understanding Information (70), the links recently placed on all Contextual Information units (6); • finds, by comparing the links (74) of the coded sets (41) chosen links (74) reactivated previous words on all Contextual Information units (6), that the choice is acceptable; • note that the sequence of words chosen forms an acceptable sentence proposal; • considers this sentence proposal properly understood; The supervisory and control system unit (2) keeps, on the Internal Projection of the Comprehension information (70), all the reactivated links recently placed on all the Contextual Information units (6), and takes over the understanding of previous proposals, providing simultaneously: • the internal display of the text understood: it rains ropes, I will go out later, warn my daughter; • the internal sequence of images: heavy rain, the hour passes, the person waiting to go outside, a representation of the device that calls Anne, and shows him the previous images; • the internal pronunciation of the text understood: it rains ropes, I will leave later, warn my daughter .. The unit Supervision and Control system (2) prepares a message for Anne on the Visual Actions of the device (75) or on the Sound Actions of the device (76) and constitutes the links (74) that correspond. This message is sent coded on the visual action parts (35) or the sound actions (36) and the links (34) of the information element time support units (30) of the units (10) of all the units Immediate Information (5) concerned. The visual actions (35) and the sound actions (36) are received by the unit (1B) which communicates the output actions. The supervisory and control system unit (2) dissociates the information units (10) containing coded sets (41) involved in the understanding of the expression from the set of Immediate Information units (5) to link them to all Contextual Information units (6), and allocates new Information units (10) to all Immediate Information units (5). The grammatical correction of sentences is taken into account but grammatically incorrect sentences can be accepted. The device (1) can even, in some cases context, help people who express themselves with difficulty: the understanding of their sentences can be facilitated by information corresponding to visual perceptions (gestures, movements, ...), and the device (1) can reformulate sentences correctly for educational purposes.

By keeping the operating bases of the present device (1), an apparatus provided with an extension of the device (1) could accept input information (1A) which is not limited to information corresponding to human visual perceptions or sound; and could provide output information (1B) that is not limited to visual or sound actions of the device (1). Thanks to its ability to identify and understand a wide range of visual and sound information, this device can • make the use of devices more natural, more immediate, more accurate; • facilitate requests for documentation and requests for services; • simplify programming, the use of computers; • serve as a language translator; • allow the commercialization of devices or robots in the general public domain to assist, protect, educate people, accompany single people; • enable the commercialization of assistive devices or robots in public services in specialized areas, such as medicine and security.

Claims (37)

An apparatus (1) for understanding human language in general comprising a) a supervisory and control system unit (2); b) an understanding unit (3); c) a set of Prior Information units (4). 2. Device according to claim 1, characterized in that the supervision and control system unit (2) manages and controls a) the operation of the device as a whole; (b) electrical, luminous, chemical energy. 3. Device according to claim 1, characterized in that the set of prior information units (4) comprises a variable number of the information unit (10). 4. Device according to claim 1, characterized in that the comprehension unit (3) comprises a) the set of Immediate Information units (5); b) all Contextual Information Units (6); c) the mechanism of Representation and Internal projection of Perceptions and Actions information (7). 20 5. Device according to claim 4, characterized in that the set of Immediate Information units (5) comprises a variable number of the information unit (10). 6. Device according to claim 4, characterized in that the set of Contextual Information units (6) comprises a variable number of the information unit (10). 7. Device according to claims 3, 5, 6, characterized in that the information unit (10) comprises a) the control system (20); b) a variable number of the information element temporal support unit 30 (30). 24 8. Device according to claim 7, characterized in that the information element temporal support unit (30) comprises: a) the information part corresponding to visual perceptions (31); (b) the information portion corresponding to sound perceptions (32); c) the part of information corresponding to the links of the Understanding (34); d) the information portion corresponding to visual actions of the device (35); e) the information portion corresponding to sound actions of the device (36). Device according to claim 8, characterized in that the information portions (31), (32), (35), (36) of the information element time support unit (30) can contain a) elementary information signals (40); b) spatiotemporal coded sets (41) consisting of elementary information information signals in connection with: • either in space and on the same information element time support unit (30); • either in space and on several units of Time Support of information elements (30). 10. Device according to claim 4, characterized in that the mechanism of Internal Representation and Projection of the information of Perceptions and Actions (7) comprises a) the Internal Representation of the information corresponding to Visual Perceptions (71), to Perceptions. Sounds (72), at Visual Actions of the device (75), at Sound Actions of the device (76); b) Internal projection of the Understanding information (70); C) the Internal Representation of the Understanding Links (74). 11. Device according to the preceding claims, characterized in that it receives, from a unit belonging to the apparatus on which is mounted the device, input sources (1A) whose information corresponds to a) to human perceptions visual; 30 b) to sound human perceptions. Device according to the preceding claims, characterized in that the set of prior information units (4) comprises, in the parts (31), (32), (35), (36), (34), of its information element temporal support units (30), information initialized to the construction of the device, respectively corresponding to human visual perceptions, to human sound perceptions, to visual actions of the device, to sound actions from the device, to the links of understanding, which may include, with varying levels, one or more dictionaries, one or more grammars, logical rules, mathematical operations, mathematical functions, specialized information. Device according to the preceding claims, characterized in that the supervisory and control system unit (2) concentrates its energy to a) chronologically decompose the input sources (1A) into elementary information signals (40), on the information portions (31), (32) of the information element time support units (30) of the information units (10) of the set of immediate information units (5); b) interconnecting the elementary information signals (40) which are found to be close by their characteristics; c) forming, from the elementary information signals (40), connected according to item b) of this claim, coded sets (41) spatio-temporal, either in space on the same unit of temporal support of information elements (30), either in space and time on several sequential units of information element temporal support (30). 14. Device according to claim 13, characterized in that the coded sets (41) correspond to the following specificities: a) a visual frame corresponding to what a human being can see in a unit of time, a sound frame corresponding to this that a human being can hear in a unit of time, can decompose into several codified sets (41); b) a visual action frame corresponding to visual actions that the device can perform in a unit of time, a sound frame of action corresponding to sound actions that the device can perform in a unit of time can be broken down into several codified sets (41); c) a sequence of information, corresponding to a chronological sequence in which visual frames, sound frames, visual action frames, sound action frames can be blended, can be broken down into several units of temporal support. information elements (30) in several coded sets (41); d) coded sets (41) can themselves be combined to form new coded sets (41); e) a subset of elementary information signals (40) of a coded set (41) can form a new coded set (41); f) a syllable, a sound, a voice, a word, a form, a criterion, a word, a diagram, a movement, a look, an object, an approach, a person, a concept, an expression, a sentence, a rule, a route, an apartment, a village, are examples of codified sets (41); (g) the same word, object, person, or concept may be defined by multiple codified sets (41); h) a word can be defined by codified sets (41) visual, sound, visual actions, sound actions, including display, listening, writing, pronunciation, inflections and variations as the gender, number, conjugation, elision, connections are examples; i) the coded sets (41) are associated with sequences of other coded sets (41), which can determine their meanings, their use cases, their grammatical uses; j) some adjacent sequences of coded sets (41), which reflect the circumstances in which codified sets or uses of sets (41) have taken place, may participate in the search for these latter coded sets (41) to which they have been associated; k) some coded sets (41) may correspond to markers of the language, including grammatical markers, logical operators, mathematical operators, grammar rules, logic rules, mathematical functions are examples. 15. Device according to the preceding claims, characterized in that the set of Contextual Information units (6) contains codified sets (41) relating to understanding processes recently performed, which may relate to events in progress, recent events, up-to-date information such as location, date. 27 Device according to the preceding claims, characterized in that the set of the prior information units (4) contain the codified sets (41) which have been preserved after having been placed beforehand, during an earlier understanding process, on these units when the latter units a) first belonged to all the Immediate Information units (5), b) and then belonged to all Contextual Information units (6). Device according to the preceding claims, characterized in that the coded sets (41) on the set of Immediate Information units (5) are made visible simultaneously by all the Contextual Information units (6). and by all the units of Prior Information (4), by an electrical, luminous, chemical process. Device according to the preceding claims, characterized in that the supervisory and control system unit (2) initiates searches, relating to the codified sets (41) made visible, which take place in parallel on the information units ( 10) of all the Contextual Information units (6) and on the information units (10) of all the Previous Information units (4), under control of the control systems (20) specific to these information units (1 0). 19. Device according to claim 18, characterized in that the searches can be a) identifying searches that take place as and when the codified sets (41); b) more specific searches taking into account or not taking into account variable criteria such as • functions, of which the identity, the homothety, the symmetry, the stretching of forms, are examples; • Boolean or logical or mathematical operations, of which yes, no, equal, unequal, and, or inclusive, or exclusive, are examples; • properties of coded sets (41), including colors, shapes, movements, steps, orientations, voices, voice intensities, accents, lyrics, pronunciation mistakes, are examples . 20. Device according to claims 18 and 19, characterized in that the searches are flexible searches, where a) a requested criterion can go from the value corresponds perfectly to the value does not correspond at all, with variable intermediate values of correspondence partial; b) the requested criteria can go from the number all the criteria correspond 5 to the number no criterion corresponds, with a variable intermediate number of criteria that correspond. Device according to the preceding claims, characterized in that the control systems (20) of the information units (10) of the set of Contextual Information units (6) and all the information units Anterior (4) may select a) the information element temporal support units (30) containing coded sets (41) previously encountered, satisfying the requested searches; b) the information element temporal support units (30) containing the coded sets (41) which have been associated with the coded sets (41) according to item a) of this claim, in an earlier process of comprehension. Apparatus according to claim 21, characterized in that the selected information element (30) temporal support units are related to the units of the immediate information set (5) where the information elements are located. the coded sets (41) to be identified or understood, by connection links established under the control of the control systems (20) specific to the information units (10) which contain them. 23. Device according to claim 22, characterized in that the connection links established a) are electrical, light, chemical; (b) vary in intensity, from very strong to very low, depending on the results of research that goes from very positive to very little positive; c) highlighting the coded sets (41) satisfying the requested searches and the coded sets (41) associated with them; D) differentiate the search results into separate groups of Time Support Units of information elements (30). 29 24. Device according to the preceding claims, characterized in that the examination of the results of the searches, under control of the Supervision and Control system unit (2) privileges first the groups of units of Temporary Support Elements information (30) a) having stronger connection links; b) having coded sets (41) that are found in multiple occurrences; c) having an acceptable combination of coded sets (41) corresponding to language markers. 25. Device according to the preceding claims, characterized in that the groups of information element temporal support units (30) chosen after examination of the search results are placed on the mechanism of representation and internal projection of the information of the information elements. Perceptions and Actions (7), under control of the Supervision and Control system unit (2), where a) the pieces of information corresponding to visual perceptions (31), to sound perceptions (32), to visual actions of the device (35), sound actions of the device (36), the links of the Understanding (34), b) are respectively placed on the Internal Representation of information corresponding to Visual Perceptions (71), Sound Perceptions (72), Visual Actions of the device (75), Sound Actions of the device (76), 20 on the Internal Representation of the Comprehension Links (74). 26. Device according to the preceding claims, characterized in that the mechanism of internal representation and projection of the information of Perceptions and Actions (7), is an internal mechanism sensitive to electrical, light, chemical reactions, which, under the control of supervisory and control system unit (2), projects in an internal space the spatio-temporal information of the selected coded sets (41) and associated coded set sequences (41); which restores their orientation, their positions, their relief; who reproduces and animates them in terms of images and sounds; who creates new associations; which represents the device itself as being an active observer of the information connections 30 constructed in its internal universe. 27. Device according to the preceding claims, characterized in that the mechanism of Internal Representation and Projection of the information of Perceptions and Actions 30 (7), exploits the Internal Representation of the Comprehension Links (74) corresponding to the links of Understanding (34). ) that have been codified in a previous understanding process. 28. Device according to the preceding claims, characterized in that, from the representation and internal projection mechanism of the Perceptions and Actions information (7), the supervision and control system unit (2) extracts or creates new data. coded sets (41), places them on units of the set of Immediate Information (5), makes them visible by the units of all Contextual Information units (6) and all the units of Previous Information (4), and launches new flexible searches, which may be identification searches or more specific searches, such as: a) searches, in all Contextual Information Units (6), of contextual values corresponding to coded sets (41) extracted; b) searching, in the set of Prior Information units (4), coded sets (41) models associated with the operation of one or more extracted language markers, which specify the rules for replacing model values with present values; c) searches, in the set of Prior Information units (4), of complementary coded sets (41), from the creation of new coded sets (41), which specify variable elements, aggregate sequences, create new internal associations. Device according to the preceding claims, characterized in that, under the control of the supervisory and control system unit (2), the information element time support units (30) of the information units ( 10) of all 25 Immediate Information units (5) containing the coded sets (41) to be identified or understood, are placed on the mechanism of Representation and Internal projection of the information of Perceptions and Actions (7), where a) the portions of information corresponding to visual perceptions (31), sound perceptions (32), visual actions of the device (35), sound actions of the device (36), links to the Comprehension (34), b) are respectively placed on the Internal Representation of information corresponding to Visual Perceptions (71), Sound Perceptions (72), 31 Visual Actions of the device (75), to Sound Actions of the device. (7 6), on the Internal Representation of Comprehension Links (74). Apparatus according to the preceding claims, characterized in that, under the control of the supervisory and control system unit (2), the internal projection of the information of understanding (70) of the mechanism of representation and internal projection of the information of Perceptions and Actions (7), a) animates internal visual sequences, internally displays the information to be identified or understood, internally displays the information identified or understood, from information corresponding to Visual Perceptions (71) or Visual Actions of the device (75); b) animates internal sound sequences, internally pronounces the information to be identified or understood, internally pronounces the information identified or understood, from the information corresponding to Sound Perceptions (72) or Sound Actions of the device ( 76); c) links together, on the Internal Representation of Comprehension Links (74), codified sets (41) relating to identification or comprehension; d) places, in coded form, the links according to the previous point c) on the link information part of the Comprehension (34) of the units (30) of the set (5). 31. Device according to the preceding claims, characterized in that the supervision and control system unit (2) controls the comprehension process by comparing, on the internal projection of the understanding information (70) of the representation and projection mechanism. internal information of Perceptions and Actions (7), a) the links that have just been created on the Internal Representation of Understanding Links (74), from the coded sets (41) of the current sequence of information to be identified or the current sentence proposal to be understood; b) the links that have just been created on the Internal Representation of the Comprehension Links (74), from the coded sets (41) chosen to identify the current information sequence or to understand the current sentence proposal; c) the comprehension links (74) resulting from the links of the Understanding (34) of all the Contextual Information units (6), relative to the last 32 sentences proposals, to the last sentences, to the last sequences of information , after their reactivation. 32. Device according to claim 31, characterized in that the supervision and control system unit (2), following the encounter of anomaly of understanding, or likelihood, or coherence, during the control of the comprehension process. (a) select, if any, other groups of information element temporal support units (30) that satisfy the search; b) resumes the comprehension process, from the selection of these groups of information element temporal support units (30). 33. Device according to claims 31 and 32, characterized in that the supervision and control system unit (2), following the encounter of anomaly of understanding, or of likelihood, or of coherence, during the control of the process of comprehension, takes again the process of comprehension, since the constitution of the initial codified sets (41), by constituting codified sets (41) different, or by taking other criteria of search. 34. Device according to claim 33, characterized in that, after resuming the comprehension process, the supervision and control system unit (2) encounters an anomaly of comprehension, or likelihood, or coherence, a) it positions coded sets (41) of action request corresponding to the type of anomaly; b) it links these coded sets (41) to the Understanding links (34). Apparatus according to the preceding claims, characterized in that, when from the Comprehension links (34) and coded sets (41) associated with them, the Supervision and Control system unit (2) detects a request for action or answer, a) it initiates searches in all Contextual Information Units (6) or in all Previous Information Units (4); b) it places the results on the mechanism of Representation and Internal projection of Perceptions and Actions information (7); c) it selects the coded sets (41) relating to the meaning of the answer or relating to the communication of the action on the Internal Representation of Understanding Links (74) and then on the Internal Representation of Information 33 corresponding to Visual Actions the device (75) or sound actions of the device (76); d) transforming the preceding information into information coded on the information portion corresponding to visual actions of the device (35) and sound actions of the device (36) of the information element temporal support units (30). information units (10) involved; e) it updates the corresponding Comprehension Links (34); f) it makes, either immediately or on the basis of events, the output actions information, thus prepared on all the immediate information units (5), available to the unit (1B) of the apparatus on which the device is mounted. 36. Device according to the preceding claims, characterized in that, when a sentence proposal or a sequence of information is considered to be included or identified by the supervision and control system unit (2), a) the units of Information (10) of all the Immediate Information units (5) containing the coded sets (41) and the Understanding links (34) associated therewith are dissociated from all the Immediate Information units (5). ) to be part of all Contextual Information Units (6); b) and new information units (10) are allocated to all Immediate Information units (5). 37. Device according to the preceding claims, characterized in that, at variable intervals, under the control of the supervision and control system unit (2), a) the information units (10) of all the units Contextual Information (6) is dissociated from the Understanding unit (3) to form part of the set of Prior information units (4); b) the information units (10) in point a) of this claim retain a variable part of their information, such as • the new coded sets (41), • the new coded sets (41) associated, • the coded sets ( 41) connected by the bonds of Understanding (34).