Pratique du Shadowing: AI Agent Memory Architecture: Structured SQL, Semantic Vector Search & Memory Bank - Apprendre l'anglais à l'oral avec la vidéo
Choisissez un mode d'entraînement
Chaque mode ouvre une salle d'entraînement en plein écran.
173 phrases
Si les phrases sont trop courtes ou trop longues, cliquez sur Edit pour les ajuster.
1
Hi, I'm Annie.
2
If your agent can hear, see, and take action, the next question is, what should I remember after the session ends?
3
In this episode, I will show you the architecture for long -term memory in a multi -model agent,
4
facts in structure memory, understanding in semantic memory, and raw audio or video outside of memory.
5
Hi there.
6
Hey there.
7
Good for that customer dinner tomorrow
8
or just browsing oh i want
9
you suggest something for customer dinner tomorrow based on my wish list for sure you have those leather ankle boots
10
and a classic black pump on your list either would go great with
11
that red cocktail dress you bought but
12
if you want something a bit more relaxed for dinner maybe mix it up with the plaid button -up
13
and classic denim jeans all right let's start with the two types of the memory when you're talking to a multimodal agent.
14
It feels like the agent remembers everything.
15
It remembers what you said five minutes ago.
16
It can use what it just heard.
17
But most of that is short -term memory.
18
It is like RAM.
19
It is useful while the session is running, but it's gone when the session ends.
20
So the real design question is not, can we save everything?
21
The real question is, what is worth carrying into the next session?
22
And And that memory comes in two kinds.
23
The first kind is facts.
24
Facts are exact.
25
You have a dinner on Thursday.
26
You bought the blue jacket.
27
Those are things you can put in rows and columns.
28
They have IDs.
29
They have dates.
30
They can be created, updated, and deleted.
31
That is structured memory.
32
The second kind is understanding.
33
Understanding is not one clean role in a table.
34
It is an age in learning what something means about you.
35
You prefer simple over flashy, you run warm, you reach for dark colors when you want to feel confident.
36
That is not just a fact to look up.
37
That is a pattern.
38
It is a difference between knowing your calendar and knowing you.
39
So, here is the first takeaway.
40
Facts are exact.
41
You look them up.
42
Understanding is meaning.
43
You recall it.
44
Now let's talk about the decision that makes or breaks your memory system.
45
Those two kinds of memory should not live in the same place.
46
Facts belong in a structured database.
47
That usually means SQL, Postgres, MySQL, Cloud SQL, and Google Cloud whatever fits your stack.
48
Why SQL?
49
Because facts need exact answers.
50
If I ask what is on my calendar Thursday, I do not want a similar memory.
51
I want an exact event.
52
The database should return the actual roles.
53
That is what structured databases are good at.
54
Now, understanding belongs somewhere else.
55
Understanding stays in a vector store.
56
A vector store lets you search by meaning.
57
So instead of asking, find a role where preference equals dark colors,
58
the agent can ask, what do I know about any style preference for this kind of event?
59
And the memory system can bring back memories that are semantically related.
60
And that is the job of embeddings and vector search.
61
Here is an important detail.
62
VectorStore is a category.
63
It is not one specific product.
64
You could use PG Vector inside Postgres.
65
You could use AlloyDB.
66
On Google Cloud, one managed option is agent platform memory bank.
67
Memory bank is useful because it does more than store vectors.
68
You can read past sessions, extract useful memories, and update those memories over time so
69
if you use a plain vector store you usually build
70
that extraction logic yourself if you use memory bank
71
that part is managed for you
72
but the lesson is not you must use memory bank the lesson is understanding goes in a system
73
that can search by meaning now here's a mistake many people make they hear semantic memory
74
and they put everything into the vector store.
75
Facts, preferences, calendar events, everything.
76
That sounds simple, but it creates a real problem.
77
A vector store is designed for similarity.
78
Similarity is great when the question is fuzzy, but similarity is dangerous.
79
when the answer must be exact.
80
If the user asks, what did I buy?
81
The agent should not guess from a similar memory.
82
It should curate the database.
83
If the user asks, what do I usually like for dinners?
84
Then semantic memory makes sense.
85
That is a meaning -based question.
86
So the rule is, structured facts goes in database like Cloud SQL.
87
Unstructured understanding goes in a vector store.
88
And there is one more category people often forget, raw perception.
89
The camera frames, the audio waveform, the live stream itself, that does not usually belong in memory.
90
Your agent does not need to permanently remember every pixel or every sound wave.
91
It needs the meaning you extract from them.
92
For example, she is wearing that black hat again.
93
Or that you just said this dinner is formal.
94
Those meanings can become memory.
95
The raw video and audio are separate storage decisions.
96
You might keep them for replay, audit, debugging, or training data.
97
But that is not the same thing as agent memory.
98
Agent memory should keep what matters, not every raw signal that pass through the app.
99
So here is a second takeaway.
100
Facts go in structured memory, like classical.
101
Understanding goes in a vector store, like agent, platform, memory, bank.
102
Role perception usually stays out of memory, unless you have a separate reason to store the media itself.
103
Alright, let's make this concrete.
104
How does the agent actually remember?
105
There are two important moments, the start of a session and the end of a session.
106
At the start, the agent loads memory before the conversation really begins.
107
It reads structured facts from the database, for example, upcoming occasions, wishlist items, profile details.
108
Then it searches the vector store for a relevant understanding, for example, style preference, comfort constraints, favorite colors from past conversations.
109
Then the agent gets both kinds of context, so when the user starts talking, the agent is not starting from zero,
110
it already has the facts and it already has a meaning.
111
That is why the code open works.
112
The agent can say, you have that dinner on Thursday from the structure database
113
and I know you like to keep it simple from semantic memory.
114
One answer, two memory systems.
115
Now let's talk about saving memory.
116
Facts should usually be saved immediately.
117
If the user says I Add that Blazor to my wish list, the agent should write that to the database right away.
118
That is a tool call.
119
It is structured, it is exact, and the user expects it to be saved now.
120
Understanding is different.
121
You usually do not want to update someone's long -term identity from one random sentence.
122
If I say, maybe I like yellow today, that does not mean yellow is now my whole style.
123
Understanding needs judgment.
124
It should come from stronger signals.
125
Maybe the user explains why something matters.
126
Maybe the same pattern repeats across sessions.
127
That is how you keep memory from getting noisy.
128
So instead of treating every sentence as a permanent preference, you let the memory series review the conversation and extract what is actually worth keeping.
129
Now, when does the extraction happen?
130
You have a choice.
131
You can extract memory during the conversation, after each turn, or at interverse.
132
Agent Platform Memory Bank supports this directly.
133
You can trigger memory generation mid -session, not only at the end
134
and extracting as you go has real advantages shorter exchanges often produce cleaner memories
135
and if the user drops off early you have already saved what they said
136
but there's a catch in a live multi -modal agent extraction
137
is heavier than a database write a model has to read the conversation
138
and infer meaning that takes time and a live voice agent is latency -sensitive,
139
you cannot let a slow extraction step block the audio.
140
So the rule is not always wait until the session ends.
141
The rule is keep extraction off the live path.
142
Run it asynchronously, after the turn or after the session, wherever it will not stall the conversation.
143
And in this demo, the structured facts are written immediately during the conversation.
144
The understanding is extracted off the hot path after the session and stored as semantic memory.
145
That gives you two tracks, fast synchronous writes for facts, slower asynchronous extraction for understanding,
146
and the next time the user comes back and the agent loads both again.
147
That is the loop.
148
Read at the start, write what changes, bring it forward next time.
149
Here's the third takeaway.
150
So both memory systems, when the session starts, save facts as soon as it happens,
151
save understanding after you have enough context to know what is actually worth remembering.
152
All right, let's recap.
153
If you want your AI agent to remember users across sessions, do not start by dumping everything into one type of database.
154
Start by asking, what kind of memory is this?
155
There are two main kinds.
156
Facts are exact.
157
It looks them up.
158
Understanding is meaning.
159
You recall it.
160
Facts belong in a structural database like classical.
161
Understanding belongs in a vector store.
162
Agent Platform Memory Bank is one managed option because it can help extract and maintain that semantic memory for you.
163
But the architecture works beyond one product and raw camera frames or raw audio do not automatically become memory.
164
You usually store the meaning you extract from them, not the raw signal itself.
165
That is what turns a forgetful agent into an agent that can actually pick up where it left off.
166
It knows the facts of your life and it understands what patterns that makes those facts useful.
167
That's it.
168
So what is the first thing you would want your agent to remember about you?
169
Don't forget to leave that in the comments.
170
In the next episode, we will give the agent eyes with real -time vision.
171
All right, I will see you in the next episode.
172
Bye.
173
Thank you.
Télécharger l'application
Tout pour parler couramment
Prononciation IAChaque phrase notée
Phonétique (IPA)Maîtrisez chaque son
VocabulaireEnrichissez votre lexique
Jeu de vocabulaireApprenez en jouant
Même chaîne
B21
The No-Code Roadmap: How I Approach Prototyping
Daily ConversationUniversity Students
C16
What's new in Google Cloud in less than 1 minute
Business EnglishUniversity StudentsWorking Adults
C15
AI harness crash course
Daily ConversationUniversity Students
C11
MCP vs API: The protocol every developer needs to know
Business EnglishUniversity StudentsWorking Adults
FOR YOU
Vidéo recommandée
C11
142 phrases
Remembering and Forgetting: Crash Course Psychology #14
Daily ConversationUniversity Students
C11
124 phrases
Agentic AI Frameworks Explained: Workflows, Multi-Agent, & Production
Daily ConversationUniversity StudentsWorking Adults
C11
389 phrases
They JUST Accidentally Revealed the Truth: The AI Bubble
Daily ConversationUniversity StudentsWorking Adults
B21
375 phrases
Configure storage accounts | AZ-104 | Episode 16
Daily ConversationUniversity StudentsWorking Adults
C11
198 phrases
Jack Dorsey’s New Dessa AI Voice Sounds Scary Human
Daily ConversationUniversity StudentsWorking Adults
![[TALK 13] Structure Prediction and Design using AlphaFold – Sami Chaaban](https://img.youtube.com/vi/my6uceci0S4/mqdefault.jpg)
C12
606 phrases
[TALK 13] Structure Prediction and Design using AlphaFold – Sami Chaaban
Daily ConversationUniversity Students
C11
712 phrases
Master AI in 60 Minutes with Vector Databases!
Daily ConversationUniversity Students
C14
23 phrases
3-D Brain Scans | Pathways 2
Daily ConversationUniversity Students
C18
55 phrases
When AIs act emotional
Daily ConversationUniversity StudentsWorking Adults
C11
345 phrases
Andrew Ng on Coding Agents and Deciding What to Build
Daily ConversationUniversity Students
C11
207 phrases
Your AI Agent Starts Every Session Like a New Hire - Killian Carlsen-Phelan, Sonar
Daily ConversationUniversity StudentsWorking Adults
C11
321 phrases
Mixture of Experts Explained Visually: How Trillion-Parameter Models Actually Work
Daily ConversationUniversity Students
Vocabulaire et conseils d’expression pour cette leçon
Cette leçon d’expression orale s’appuie sur la vidéo « AI Agent Memory Architecture: Structured SQL, Semantic Vector Search & Memory Bank ». Les mots qui reviennent le plus souvent : memory, agent, understanding, session, vector. Cette vidéo contient 173 phrases et 1681 mots à répéter en shadowing. La partie parlée dure 12:17. Le locuteur parle à un rythme régulier d’environ 137 mots par minute, confortable pour le shadowing. 84 % des mots font partie des 3 000 mots les plus courants en anglais ; le reste mérite d’être vu avant de commencer.
Vocabulaire clé de cette vidéo
15 mots de la vidéo à retenir, avec leur prononciation et leur sens :
| Mot | Prononciation | Sens |
|---|---|---|
| vector nom | /ˈvɛktɚ/ | vecteur |
| database nom | /ˈdeɪtəˌbeɪs/ | base de données, banque de données |
| extract nom | /ˈɛkstɹækt/ | extrait |
| extraction nom | /ɪkˈstɹækʃən/ | extraction |
| audio adjectif | /ˈɔ.di.oʊ/ | audio |
| belong verbe | /bɪˈlɒŋ/ | avoir sa place |
| preference nom | /ˈpɹɛf(ə)ɹ(ə)ns/ | préférence |
| calendar nom | /ˈkæl.ən.də/ | calendrier |
| similarity nom | /ˌsɪm.ɪˈlæɹ.ɪ.ti/ | similarité, similitude |
| lesson nom | /ˈlɛs.ən/ | leçon |
| architecture nom | /ˈɑː.kɪˌtɛk.t͡ʃə/ | architecture |
| recall verbe | /ɹɪˈkɔl/ | rappeler |
| perception nom | /pɚˈsɛpʃ(ə)n/ | perception |
| sensitive adjectif | /ˈsɛn.sɪt.ɪv/ | sensible |
| formal adjectif | /ˈfɔɹməl/ | formel, soutenu |
Les verbes à particule que vous entendrez
| Mot | Prononciation | Sens |
|---|---|---|
| bring back verbe | /ˈbɹɪŋ bæk/ | ramener, rapporter |
| come back verbe | revenir | |
| drop off verbe | s’endormir, tomber | |
| pass through verbe | traverser | |
| pick up verbe | soulever |
Des phrases à répéter
Des phrases courtes et complètes de la vidéo, à réutiliser dans la conversation de tous les jours :
- If the user asks, what did I buy?
- If the user asks, what do I usually like for dinners?
- Alright, let's make this concrete.
Prononciation à surveiller
Le locuteur utilise 3 contractions et formes réduites, comme don't, I'm, you're. Prononcez-les sous leur forme courte, telles que vous les entendez.
- Les sons « sh » et « zh »: extraction /ɪkˈstɹækʃən/, perception /pɚˈsɛpʃ(ə)n/, occasion /əˈkeɪ̯ʒən/, structural /ˈstɹʌk(t͡)ʃəɹəl/, flashy /ˈflæʃi/
- Mots longs — placez bien l’accent: similarity /ˌsɪm.ɪˈlæɹ.ɪ.ti/, architecture /ˈɑː.kɪˌtɛk.t͡ʃə/, permanently /ˈpɜː.mə.nənt.li/, asynchronous /eɪˈsɪŋ.kɹə.nəs/
Les sons difficiles pour les francophones :
- /tʃ/ et /dʒ/ — à ne pas adoucir en « ch » et « j »: architecture /ˈɑː.kɪˌtɛk.t͡ʃə/, jacket /ˈd͡ʒæk.ɪt/, judgment /ˈd͡ʒʌd͡ʒ.mənt/, jeans /d͡ʒiːnz/
- /r/ anglais — langue recourbée, sans frotter la gorge: extract /ˈɛkstɹækt/, extraction /ɪkˈstɹækʃən/, preference /ˈpɹɛf(ə)ɹ(ə)ns/, similarity /ˌsɪm.ɪˈlæɹ.ɪ.ti/, recall /ɹɪˈkɔl/
- /ŋ/ final — sans prononcer de /g/ après: belong /bɪˈlɒŋ/, embedding /ɪmˈbɛdɪŋ/
Comment s’entraîner avec cette vidéo
- Écoutez la vidéo en entier une fois sans parler et notez les mots que vous ne connaissez pas.
- Répétez phrase par phrase à vitesse normale, en reprenant chacune jusqu’à ce que votre rythme corresponde à celui du locuteur.
- Enregistrez-vous et comparez avec l’original, en faisant attention à des mots comme vector, database, extract.
Qu'est-ce que la technique du Shadowing ?
Le Shadowing est une technique d'apprentissage des langues fondée sur la science, développée à l'origine pour la formation des interprètes professionnels. Le principe est simple mais puissant : vous écoutez de l'anglais natif et le répétez immédiatement à voix haute — comme une ombre suivant le locuteur avec un décalage de 1 à 2 secondes. Les recherches montrent une amélioration significative de la précision de la prononciation, de l'intonation, du rythme, des liaisons, de la compréhension orale et de la fluidité.