This is an automated email from the ASF dual-hosted git repository.
aradzinski pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-nlpcraft-website.git
The following commit(s) were added to refs/heads/master by this push:
new 4b2527f WIP.
4b2527f is described below
commit 4b2527f18071f618e3743904d352e228ad1f7d04
Author: Aaron Radzinski <aradizn...@apache.org>
AuthorDate: Wed Jul 28 13:20:22 2021 -0700
WIP.
---
_includes/latest_ver_blog_warn.html | 2 +-
_scss/misc.scss | 4 +-
_scss/short-term-memory.scss | 11 ++-
intent-matching.html | 8 ++-
short-term-memory.html | 130 +++++++++++++++---------------------
5 files changed, 73 insertions(+), 82 deletions(-)
diff --git a/_includes/latest_ver_blog_warn.html
b/_includes/latest_ver_blog_warn.html
index c46cfed..d6a99bc 100644
--- a/_includes/latest_ver_blog_warn.html
+++ b/_includes/latest_ver_blog_warn.html
@@ -20,6 +20,6 @@
<p>
At the moment of this writing, the project undergoes incubation at <a
target=asf href="https://apache.org/";>Apache Software Foundation</a>, the
<a href="/download.html">latest version</a> of NLPCraft is
<b>{{site.latest_version}}</b>. In some rare cases, APIs might have changed
- since the this article was written. Consult the latest <a
href="/docs.html">documentation</a> for up to date information.
+ since this article was written. Consult the latest <a
href="/docs.html">documentation</a> for up-to-date information.
</p>
</div>
\ No newline at end of file
diff --git a/_scss/misc.scss b/_scss/misc.scss
index 612744b..ca433a1 100644
--- a/_scss/misc.scss
+++ b/_scss/misc.scss
@@ -357,6 +357,8 @@ $bq-success-border-color: $brand-success;
.max-width {
width: 100%;
}
+
+.not-code {}
a[target=javadoc]:not(.not-code) {
font-family: SFMono-Regular, Menlo, Monaco, Consolas, "Liberation Mono",
"Courier New", monospace;
word-break: keep-all;
@@ -370,7 +372,7 @@ pre.console {
}
code {
- font-size: 100%;
+ font-size: 105%;
white-space: nowrap !important;
color: #C0392B;
diff --git a/_scss/short-term-memory.scss b/_scss/short-term-memory.scss
index e73387d..e6ca988 100644
--- a/_scss/short-term-memory.scss
+++ b/_scss/short-term-memory.scss
@@ -52,15 +52,24 @@
display: block;
}
+ i {
+ font-size: 90%;
+ color: #ccc;
+ }
+
span {
+ vertical-align: top;
min-width: 50px;
display: inline-block;
padding: 3px 6px;
border: 1px solid #C5D2F7 !important;
border-radius: 2px;
text-align: center;
- font-weight: 200;
letter-spacing: 0.02em;
+
+ b {
+ color: $color-wisteria;
+ }
}
}
}
diff --git a/intent-matching.html b/intent-matching.html
index 4dc8f3f..f39d14d 100644
--- a/intent-matching.html
+++ b/intent-matching.html
@@ -182,6 +182,7 @@ id: intent_matching
<thead>
<tr>
<td>Option</td>
+ <td>Type</td>
<td>Description</td>
<td>Default Value</td>
</tr>
@@ -189,6 +190,7 @@ id: intent_matching
<tbody>
<tr>
<td><code>ordered</code></td>
+ <td><code>Boolean</code></td>
<td>
Whether or not this intent is ordered.
For ordered intent the specified order
of terms is important for matching this intent.
@@ -201,6 +203,7 @@ id: intent_matching
</tr>
<tr>
<td><code>unused_free_words</code></td>
+ <td><code>Boolean</code></td>
<td>
Whether or not free words - that are
unused by intent matching - should be
ignored (value <code>true</code>) or
reject the intent match (value <code>false</code>).
@@ -212,6 +215,7 @@ id: intent_matching
</tr>
<tr>
<td><code>unused_sys_toks</code></td>
+ <td><code>Boolean</code></td>
<td>
Whether or not unused <a
href="/data-model.html#builtin">system tokens</a> should be
ignored (value <code>true</code>) or
reject the intent match (value <code>false</code>).
@@ -221,16 +225,18 @@ id: intent_matching
</tr>
<tr>
<td><code>unused_usr_toks</code></td>
+ <td><code>Boolean</code></td>
<td>
Whether or not unused user-defined
tokens should be
ignored (value <code>true</code>) or
reject the intent match (value <code>false</code>).
By default, tne unused user tokens are
not ignored since it is assumed that user would
- defined his or her own tokens on
purpose and construct the intent logic appropriate.
+ define his or her own tokens on
purpose and construct the intent logic appropriate.
</td>
<td><code>false</code></td>
</tr>
<tr>
<td><code>allow_stm_only</code></td>
+ <td><code>Boolean</code></td>
<td>
Whether or not the intent can match
when all of the matching tokens came from STM.
By default, this special case is
disabled (value <code>false</code>). However, in specific intents
diff --git a/short-term-memory.html b/short-term-memory.html
index 255294b..c70973d 100644
--- a/short-term-memory.html
+++ b/short-term-memory.html
@@ -22,16 +22,18 @@ id: short_term_memory
-->
<div id="short-term-memory" class="col-md-8 second-column">
- <section id="stm">
+ <section id="overview">
<h2 class="section-title">Conversation <span class="amp">&</span> STM
<a href="#"><i class="top-link fas fa-fw fa-angle-double-up"></i></a></h2>
<p>
NLPCraft provides automatic conversation management right out of
the box that is fully integrated with
- <a href="/intent-matching.html">intent matching</a>. Conversation
management is based on the idea of short-term memory (STM). STM is automatically
- maintained by NLPCraft per each user and data model combination.
Essentially, NLPCraft "remembers"
- the context of the conversation and can supply the currently
missing elements from STM
- to the intent matching algorithm for the conversational terms.
+ <a href="/intent-matching.html">intent matching</a>. Conversation
management is based on the idea of short-term memory (STM) -
+ the mechanism by which NLPCraft "remembers" the context of the
conversation for a certain amount of time and certain
+ depth of the conversation. STM is automatically maintained by
NLPCraft per each user and data model combination.
+ Intent matching algorithm can "recall" the missing tokens from STM
when trying to find a match for the
+ conversational terms.
</p>
<p>
+ Why is this so important?
Maintaining conversation state is necessary for effective context
resolution, so that users
could ask, for example, the following sequence of questions using
example weather model:
</p>
@@ -45,13 +47,13 @@ id: short_term_memory
</p>
<div class="stm-state">
<div class="stm">
- <label>STM Before:</label>
+ <label><b>STM</b> Before:</label>
<span> </span>
</div>
<div class="stm">
- <label>STM After:</label>
+ <label><b>STM</b> After:</label>
<span>weather</span>
- <span>London</span>
+ <span>london</span>
<span>today</span>
</div>
</div>
@@ -67,15 +69,15 @@ id: short_term_memory
</p>
<div class="stm-state">
<div class="stm">
- <label>STM Before:</label>
+ <label><b>STM</b> Before:</label>
<span>weather</span>
<span>London</span>
<span>today</span>
</div>
<div class="stm">
- <label>STM After:</label>
+ <label><b>STM</b> After:</label>
<span>weather</span>
- <span><b>Berlin</b></span>
+ <span><b>berlin</b><br/><i>london</i></span>
<span>today</span>
</div>
</div>
@@ -90,16 +92,16 @@ id: short_term_memory
</p>
<div class="stm-state">
<div class="stm">
- <label>STM Before:</label>
+ <label><b>STM</b> Before:</label>
<span>weather</span>
<span>Berlin</span>
<span>today</span>
</div>
<div class="stm">
- <label>STM After:</label>
- <span><b>forecast</b></span>
- <span>Berlin</span>
- <span><b>Next week</b></span>
+ <label><b>STM</b> After:</label>
+ <span><b>forecast</b><br/><i>weather</i></span>
+ <span>berlin</span>
+ <span><b>next week</b><br/><i>today</i></span>
</div>
</div>
</dt>
@@ -109,24 +111,14 @@ id: short_term_memory
Conversation management implementation is also smart enough to
clear STM after certain
period of time, i.e. it “forgets” the conversational context after
few minutes of inactivity.
Note also that conversational context can also be cleared
explicitly
- via <a
href="https://github.com/apache/incubator-nlpcraft/blob/master/openapi/nlpcraft_swagger.yml";
target="github">REST API</a>.
+ via <a
href="https://app.swaggerhub.com/apis-docs/Apache-NLPCraft/apache-nlpcraft/{{site.latest_version}}";
target="swaggerhub">REST API</a>
+ or from the <a class="not-code" target="javadoc"
href="/apis/latest/org/apache/nlpcraft/model/NCConversation.html">code</a>.
</p>
- </section>
- <section>
- {% include latest_ver_blog_warn.html %}
<p>
- In this blog, I'll try to give a high-level overview of STM -
Short-Term Memory, a technique used to
- maintain conversational context in NLPCraft. Maintaining the
proper conversation context - remembering
- what the current conversation is about - is essential for all
human interaction and thus essential for
- computer-based natural language understanding. To my knowledge,
NLPCraft provides one of the most advanced
- implementations of STM, especially considering how tightly it is
integrated with NLPCraft's unique
- intent-based matching (Google's <a target=google
href="https://cloud.google.com/dialogflow/";>DialogFlow</a> is very similar yet).
- </p>
- <p>
- Let's dive in.
+ To understand the algorithm behind the STM management let's back
up a few steps...
</p>
</section>
- <section>
+ <section id="parsing">
<h2 class="section-title">Parsing User Input <a href="#"><i
class="top-link fas fa-fw fa-angle-double-up"></i></a></h2>
<p>
One of the key objectives when parsing user input sentence for
Natural Language Understanding (NLU) is to
@@ -159,8 +151,8 @@ id: short_term_memory
sentence since without a context the missing information cannot be
obtained and the sentenced cannot be interpreted.
</p>
</section>
- <section>
- <h2 class="section-title">Semantic Entities <a href="#"><i
class="top-link fas fa-fw fa-angle-double-up"></i></a></h2>
+ <section id="ne">
+ <h2 class="section-title">Named Entities <a href="#"><i
class="top-link fas fa-fw fa-angle-double-up"></i></a></h2>
<p>
Let's take a closer look at the named entities from the above
examples:
</p>
@@ -209,7 +201,7 @@ id: short_term_memory
</p>
</div>
</section>
- <section>
+ <section id="incomplete">
<h2 class="section-title">Incomplete Sentences <a href="#"><i
class="top-link fas fa-fw fa-angle-double-up"></i></a></h2>
<p>
Assuming previously asked questions about the weather in Tokyo (in
the span of the ongoing conversation) one
@@ -229,10 +221,10 @@ id: short_term_memory
</ul>
<p>
These are incomplete sentences. This type of short-hands cannot be
interpreted without prior context (neither
- by humans or by machines) since by themselves they are missing
necessary information.
+ by humans nor by machines) since by themselves they are missing
necessary information.
In the context of the conversation, however, these incomplete
sentences work. We can simply provide one or two
entities and rely on the <em>"listener"</em> to recall the rest of
missing information from a
- <em>short-term memory</em>, a.k.a conversation context.
+ short-term memory.
</p>
<p>
In NLPCraft, the intent-matching logic will automatically try to
find missing information in the
@@ -241,14 +233,14 @@ id: short_term_memory
found explicitly in the user input.
</p>
</section>
- <section>
+ <section id="stm">
<h2 class="section-title">Short-Term Memory <a href="#"><i
class="top-link fas fa-fw fa-angle-double-up"></i></a></h2>
<p>
The short-term memory is exactly that... a memory that keeps only
small amount of recently used information
and that evicts its contents after a short period of inactivity.
</p>
<p>
- Let's look at the example from a real life. If you would call your
friend in a couple of hours asking <code>"What about a day after?"</code>
+ Let's look at the example from a real life. If you called your
friend in a couple of hours asking <code>"What about a day after?"</code>
(still talking about weather in Kyoto) - he'll likely be
thoroughly confused. The conversation is timed out, and
your friend has lost (forgotten) its context. You will have to
explain again to your confused friend what is that you are asking about...
</p>
@@ -257,7 +249,7 @@ id: short_term_memory
what happens if we switch the topic before this timeout elapsed?
</p>
</section>
- <section>
+ <section id="ctx-switch">
<h2 class="section-title">Context Switch <a href="#"><i
class="top-link fas fa-fw fa-angle-double-up"></i></a></h2>
<p>
Resetting the context by the timeout is, obviously, not a hard
thing to do. What can be trickier is to detect
@@ -270,7 +262,7 @@ id: short_term_memory
</p>
<ul>
<li>
- <code>"How much mocha latter at Starbucks?"</code><br/>
+ <code>"How much is mocha latter at Starbucks?"</code><br/>
At this point we should forget all about previous conversation
about weather and assume going forward
that we are talking about coffee in Starbucks.
</li>
@@ -287,11 +279,16 @@ id: short_term_memory
Despite somewhat obvious logic the implementation of context
switch is not an exact science. Sometimes, you
can have a "soft" context switch where you don't change the topic
of the conversation 100% but yet sufficiently
enough to forget at least some parts of the previously collected
context. NLPCraft has a built-in algorithm
- to detect the hard switch in the conversation. It also exposes API
to perform a selective reset on the
+ to detect the hard switch in the conversation. It also exposes
+ <a class="not-code" target="javadoc"
href="/apis/latest/org/apache/nlpcraft/model/NCConversation.html">API</a> to
perform a selective reset on the
conversation in case of "soft" switch.
</p>
+ <p>
+ See <a class="javadoc"
href="/apis/latest/org/apache/nlpcraft/model/NCConversation.html">NCConversion</a>
interface
+ for API details for STM management.
+ </p>
</section>
- <section>
+ <section id="override">
<h2 class="section-title">Overriding Entities <a href="#"><i
class="top-link fas fa-fw fa-angle-double-up"></i></a></h2>
<p>
As we've seen above one named entity can replace or override an
older entity in the STM, e.g. <code>"Peet's"</code>
@@ -348,7 +345,7 @@ id: short_term_memory
</li>
</ul>
</section>
- <section>
+ <section id="rule">
<h2 class="section-title">Overriding Rule <a href="#"><i
class="top-link fas fa-fw fa-angle-double-up"></i></a></h2>
<p>
Here's the rule we developed at NLPCraft and have been
successfully using in various models:
@@ -440,12 +437,8 @@ id: short_term_memory
</ul>
</li>
</ul>
- <p>
- And so on... easy, huh 😇 In fact, the logic is indeed relatively
straightforward. It also follows
- common sense where the logic produced by this rule matches the
expected human behavior.
- </p>
</section>
- <section>
+ <section id="switch">
<h2 class="section-title">Explicit Context Switch <a href="#"><i
class="top-link fas fa-fw fa-angle-double-up"></i></a></h2>
<p>
In some cases you may need to explicitly clear the conversation
STM without relying on algorithmic behavior.
@@ -474,43 +467,24 @@ id: short_term_memory
in the NLPCraft model.
</p>
<p>
- In NLPCraft you can also explicitly reset conversation context
through API or by switching the model on the request.
- </p>
- </section>
- <section>
- <h2 class="section-title">Summary <a href="#"><i class="top-link fas
fa-fw fa-angle-double-up"></i></a></h2>
- <p>
- Let’s collect all our thoughts on STM into a few bullet points:
+ In NLPCraft you can also explicitly reset conversation context
through
+ <a class="not-code" target="javadoc"
href="/apis/latest/org/apache/nlpcraft/model/NCConversation.html">API</a>
+ or by switching the model on the request.
</p>
- <ul>
- <li>
- Missing entities in incomplete sentences can be auto-recalled
from STM.
- </li>
- <li>
- Newly detected type/category entity is likely indicating the
change of topic.
- </li>
- <li>
- The key property of STM is its short-time storage and
overriding rule.
- </li>
- <li>
- The explicit context switch is an important mechanism.
- </li>
- </ul>
- <div class="bq info">
- <b>
- Short-Term Memory
- </b>
- <p>
- It is uncanny how properly implemented STM can make
conversational interface <b>feel like a normal human
- conversation</b>. It allows to minimize the amount of
parasitic dialogs and Q&A driven interfaces
- without unnecessarily complicating the implementation of such
systems.
- </p>
- </div>
</section>
</div>
<div class="col-md-2 third-column">
<ul class="side-nav">
<li class="side-nav-title">On This Page</li>
+ <li><a href="#overview">Conversation <span class="amp">&</span>
STM</a></li>
+ <li><a href="#parsing">Parsing User Input</a></li>
+ <li><a href="#ne">Named Entities</a></li>
+ <li><a href="#incomplete">Incomplete Sentences</a></li>
+ <li><a href="#stm">Short-Term-Memory</a></li>
+ <li><a href="#ctx-switch">Context Switch</a></li>
+ <li><a href="#override">Overriding Entities</a></li>
+ <li><a href="#rule">Overriding Rule</a></li>
+ <li><a href="#switch">Context Switch</a></li>
{% include quick-links.html %}
</ul>
</div>
