晓之车高山老师 · Bilibili · 8:34
This segment introduces the concept of quadratic forms using univariate quadratic functions. The video first reviews and , then explains that after coordinate translation it becomes , calling this form—containing only quadratic terms—the homogeneous quadratic function. It then generalizes to the bivariate example , emphasizing that the total degree of the cross term is also 2. The core definition is: a multivariate quadratic function is homogeneous—and thus a quadratic form—if every term has degree 2; therefore, quadratic forms may contain only square terms and cross terms involving two variables. Finally, three positive examples and four counterexamples illustrate the classification method; the counterexamples , , , and +2 are excluded because their degrees are not 2. This segment explains how to convert a ternary real quadratic form into matrix form. Using a concrete polynomial as an example, the video demonstrates halving cross terms, grouping and factoring, and writing the expression as a row vector times a matrix times a column vector, ultimately yielding , where A is a real symmetric matrix. It then illustrates, via several non-symmetric matrix representations, that matrix representations are generally not unique, whereas the symmetric matrix representation is the unique standard form. This segment explains the relationship between quadratic forms and real symmetric matrices. First, using , it illustrates that a single quadratic form may have multiple matrix representations, but only the real symmetric matrix representation is unique; then it gives the direct writing rule: place coefficients of square terms on the main diagonal, and halve coefficients of cross terms before placing them symmetrically at positions a_{ij} and a_{ji}. Next, it presents two special cases: when only square terms are present, the matrix is diagonal; when only cross terms are present, all entries on the main diagonal are 0. Finally, it defines the rank of a quadratic form as the rank of its corresponding real symmetric matrix, and reminds viewers to verify real symmetry before computing the rank. This segment mainly explains the matrix representation relationships of quadratic forms under different bases. It derives the congruence transformation formula via coordinate transformation, defines congruence transformation, and points out that congruence and similarity are both equivalence relations on matrices. Finally, it introduces the research goal of reducing quadratic forms to canonical form via diagonalization. This segment explains matrix congruence and its application in quadratic form diagonalization, pointing out that a congruence transformation via an invertible matrix can convert a quadratic form into a special form containing only square terms. Meanwhile, the video raises two follow-up questions: first, the practical significance of simplifying quadratic forms; second, the relationship between congruence transformations and similarity diagonalization—laying the groundwork for deeper exploration later.
Use the learning inspector for key ideas and moments, or open the reading tabs for the complete notes.
Generated from the video's visuals and explanation; not verbatim speech.
The video opens with the lecture title 'Lecture 44: Concept of Quadratic Forms', then proceeds to blackboard instruction.
Starting from the univariate quadratic function familiar to students, the video writes on the blackboard and adds its completed-square form . The emphasis here is to display both the general form and the vertex form of a 'quadratic function' side by side, laying the groundwork for later isolating the case containing only quadratic terms.
Next, the narration explains that translating coordinates can move the parabola's vertex to the origin, thereby simplifying the expression to . The screen simultaneously writes this simplified formula and labels it beside as 'homogeneous quadratic function'. Mathematically, this step does not introduce a new operation; rather, it uses translation to eliminate the linear and constant terms, retaining only terms whose total degree is 2.
Then the video extends the perspective from univariate to multivariate cases, writing on the blackboard. Here, and are square terms, and is a cross term. The narration specially clarifies that although involves the product of two distinct variables, its total degree remains 2, so it does not violate the condition of being 'homogeneous quadratic'.
Building on this, the video gives the formal definition: a multivariate quadratic function is called homogeneous—and thus a quadratic form—if every one of its terms has degree 2. It immediately follows up with a more intuitive criterion: a quadratic form may contain only square terms and cross-product terms involving exactly two variables. In other words, verification requires checking not merely whether quadratic terms exist, but whether every single term has degree 2.
To solidify the definition, the blackboard first lists three positive examples: , , and . Each term in these three expressions is either the square of a single variable or the product of two variables, satisfying the definition of a quadratic form.
The video then lists four counterexamples, highlighting with white boxes the terms that violate the conditions: is a cubic term, is a three-variable cross product term, is a linear term, and is a constant term. The corresponding expressions are , , , and . Finally, a red cross appears on screen, emphasizing that if any non-quadratic term is present, the entire expression is not a quadratic form.
At the beginning of the video, several quadratic forms are written on a blackboard, and the voiceover introduces the topic: exploring the relationship between quadratic forms and matrices.
Next, a specific three-variable quadratic form is used as an example to explain the crucial first step—halving the cross terms. For instance, 4x₁x₂ is split into 2x₁x₂ + 2x₁x₂, so that the resulting matrix will be symmetric about its main diagonal.
Then, the halved terms are grouped by variables, and the common factors x₁, x₂, and x₃ are factored out from each group, rewriting the expression as a sum of three products: each variable multiplied by a corresponding parenthesis.
This step naturally leads to vector notation: a row vector (x₁ x₂ x₃) multiplied by a column vector formed by those parentheses. Since each entry in the column vector is a linear combination of the variables, it can be further decomposed into a coefficient matrix multiplied by the column vector of variables.
The standard matrix form ⁼TAx is finally obtained, with the special note ⁴, emphasizing that the middle coefficient matrix is symmetric.
The voiceover concludes: every real quadratic form corresponds to a real symmetric matrix. This is one of the core conclusions of this section.
The video then warns of a common misconception: without halving the cross terms, one can still write some matrix representation, but that matrix is not symmetric, and such a representation is not unique.
Finally, by comparing multiple different matrices shown on the blackboard, the conclusion is made clear: only the real symmetric matrix obtained by halving the cross terms gives the unique matrix representation of a quadratic form.
This segment first answers the question: "Can quadratic forms be studied just like matrices?" On screen, the same quadratic form is written in three different matrix product forms, and the conclusion is explicitly stated: the matrix representation of a quadratic form is not unique—it can have infinitely many representations; only when expressed as a real symmetric matrix is the representation unique. Therefore, all subsequent methods focus on the "real symmetric matrix representation."
Next, we move to the concrete writing method. The video retains the real symmetric matrix representation with a white border, and uses colored boxes and connecting lines to match each term of the polynomial with corresponding matrix entries one by one. The rule is: place the coefficient of a square term directly on the main diagonal at position ; for a cross term , halve its coefficient and place the result symmetrically at positions and . Taking as an example, its coefficient 4 is halved to yield 2, so entries at row 1 column 2 and row 2 column 1 of the matrix are both filled with 2. The same applies to all other cross terms, resulting finally in .
Then two special cases are used to reinforce the rule. The first is , which contains only square terms and no cross terms, so all off-diagonal entries are 0, and the corresponding matrix is the diagonal matrix . The second is , which contains only cross terms and no square terms, so all main diagonal entries are 0, and the coefficients of the cross terms are halved and placed symmetrically, yielding . These two examples illustrate: whether square terms are present determines whether the main diagonal entries are nonzero, and whether cross terms are present determines whether the off-diagonal entries are nonzero.
Finally, the topic shifts to "the rank of a quadratic form." The screen first returns to the general expression , annotated with , then writes down . The explanation emphasizes that the rank of a quadratic form is not defined for an arbitrary matrix representation, but rather as "the rank of the corresponding real symmetric matrix." Therefore, if a problem gives a quadratic form in some form, one must first check whether is real symmetric; if not, it must first be converted into a real symmetric matrix before computing its rank. At the end of this segment, this idea begins to be linked with previously studied concepts—basis change and similarity transformation—but no further derivations are carried out within this segment.
First, recall that the rank of a quadratic form is defined as the rank of its corresponding real symmetric matrix, namely . Note that the matrix representation of a quadratic form is not unique; thus, when handling the form , always verify whether is a real symmetric matrix.
Next, we explore the same quadratic form expressed in different bases. Suppose the quadratic form has the expression with respect to the basis vectors , where is the corresponding coordinate variable. Similarly, under another set of basis vectors , the expression is , where are the new coordinate variables.
Now we derive the relationship between these two matrices and . Let the change of basis matrix from basis to basis be , and according to the coordinate transformation formula, we have . Substituting Substitute into the first expression to get . Using the property of matrix transpose , expand to get .
Comparing with , we conclude that . This form is called matrix congruence transformation. It means that the real symmetric matrices representing the same quadratic form under different bases are congruent to each other, and the matrix realizing this transformation is precisely the change-of-basis matrix; therefore, must be invertible.
The congruence relation between matrices closely resembles the similarity relation between matrices studied earlier. Similar matrices describe the matrix representations of the same linear transformation under different bases, whereas congruent matrices describe the matrix representations of the same quadratic form under different bases. Both belong to special equivalence relations in matrix theory and hold significant research value.
The purpose of studying these equivalence relations is to simplify problems. For similar matrices, we aim to diagonalize them to simplify linear transformations; for quadratic forms, we also seek a suitable invertible matrix such that, after congruence transformation, becomes a diagonal matrix , i.e., . Thus, the quadratic form reduces to a special form containing only square terms, eliminating cross terms.
In the previous section, we learned that by choosing an appropriate invertible matrix , matrix can be transformed via congruence into a diagonal matrix , satisfying formula . This step converts a general quadratic form into a special quadratic form containing only square terms.
At this point, a natural question arises: What important practical significance or theoretical value lies in converting an arbitrary quadratic form into a form containing only square terms? This involves a deep understanding of the geometric shape and algebraic properties of quadratic forms.
Moreover, observing the congruence transformation expression , we notice its striking resemblance to another core concept in linear algebra—the similarity diagonalization formula for real symmetric matrices. Then, what intrinsic connections and essential differences exist between these two transformations? We will explore these questions in depth in subsequent lessons.
The video first reviews middle-school content: a univariate quadratic function can be written as , with explicitly labeled on screen; it can further be rewritten in completed-square form as . This step transforms the 'quadratic function' from its general form into the vertex form, which makes translational effects easier to observe.
After translating the parabola's vertex to the origin via coordinate translation, the linear and constant terms vanish, leaving only . The video calls this univariate form containing only quadratic terms a 'homogeneous quadratic function'. Its defining feature is that every term has degree 2, with no lower-degree terms present to interfere.
The video uses to illustrate the multivariate case. and are square terms, and is a cross term. Crucially, although involves two distinct variables, its total degree remains 2, so it qualifies as an allowed term in a quadratic form.
The definition given in the video is: a multivariate quadratic function is called homogeneous—and thus a quadratic form—if every term has total variable degree 2. In other words, a quadratic form is not merely 'a function containing quadratic terms', but rather 'a function whose every term must be exactly quadratic'. This includes both single-variable squares and cross products of two variables.
Complementing the definition, the video further explains: only square terms and cross-product terms involving exactly two variables are allowed in a quadratic form. Thus, linear terms, constant terms, cubic terms, and terms involving the product of three or more variables are all prohibited. This criterion facilitates quick screening of expressions.
In practice, one should check the total degree of each term individually: if every term has degree 2, then it is a quadratic form; if even one term has degree different from 2, then it is not a quadratic form. The positive and negative examples later in the video follow this method, with violating terms highlighted in boxes.
, , and are all quadratic forms because all their terms are either square terms or cross-product terms of two variables, each having degree 2.
contains the cubic term ; contains the triple-product term ; contains the linear term ; contains the constant term . As soon as any such non-quadratic term appears, the entire expression fails to be a quadratic form.
The video uses counterexamples to emphasize that determining whether an expression is a quadratic form requires checking every single term—not merely checking whether terms like or appear. Even if most terms are quadratic, the presence of just one term like , , or a constant term renders the whole expression non-quadratic.
The central question addressed in this segment is: how can a polynomial-style quadratic form be written in a concise matrix form? The video demonstrates the complete path from the polynomial to x⁼TAx using a concrete example.
To construct a symmetric matrix, each cross term must be evenly split into two parts. For example, 4x₁x₂ is split into 2x₁x₂ + 2x₁x₂. This operation is essential to ensure that the matrix A satisfies ⁴.
After grouping and factoring out common factors, a quadratic form can be written as the product of a row vector, a coefficient matrix, and a column vector. The diagonal entries of the middle matrix correspond to the coefficients of the squared terms, and the off-diagonal entries correspond to half the coefficients of the cross terms.
The video points out that if symmetry is not required, the same quadratic form may be represented by infinitely many matrices; however, once restricted to real symmetric matrices, the representation becomes unique. Therefore, textbooks and examinations default to using symmetric matrices for representation.
The same quadratic form can be written in multiple forms, so its matrix representation is not unique. The video displays three different matrices side by side for the same , illustrating that one must not mistakenly treat any single matrix representation as the "unique answer."
Among all possible matrix representations, the matrix representation of a quadratic form is unique only when the matrix is required to be real symmetric. This is also why the video stresses that "methods for studying real symmetric matrices can be applied to study quadratic forms."
When writing the real symmetric matrix, the coefficient of the square term is placed at the main diagonal position ; the coefficient of the cross term is first halved, then placed symmetrically at positions and . Using as an example, the video halves 4 to obtain 2, and places it at row 1, column 2 and at row 2, column 1.
If a quadratic form contains only square terms and no cross terms, then all off-diagonal entries are 0, and the corresponding real symmetric matrix is diagonal. The video example is .
If a quadratic form contains only cross terms and no square terms, then all main diagonal entries are 0, and the nonzero entries arise from the symmetric placement of the halved cross-term coefficients. The video example is .
If a quadratic form is written as and , then define . That is, the rank of a quadratic form is not defined for an arbitrary representing matrix, but specifically for its unique real symmetric matrix.
Because the matrix representation of a quadratic form is not unique, when encountering a form , one cannot directly take the rank of as the rank of the quadratic form. One must first check whether is real symmetric; if not, one must first convert it into a real symmetric matrix and then compute its rank.
The rank of a quadratic form is defined as the rank of its associated real symmetric matrix. Note: the matrix representation of a quadratic form is not unique; before use, confirm whether the matrix is real symmetric.
If there exists an invertible matrix P such that , then matrix B is said to be congruent to A, and this transformation is called a congruence transformation. It reflects the relationship among matrix representations of the same quadratic form under different basis vector sets.
Both congruence transformation and similarity transformation are special equivalence relations on matrices. Similarity transformation corresponds to matrix representations of the same linear transformation under different bases; congruence transformation corresponds to matrix representations of the same quadratic form under different bases. In both cases, a key research goal is to find an appropriate transformation matrix to diagonalize the matrix, thereby simplifying the problem.
Matrix congruence is an equivalence relation on matrices, reflecting how the same quadratic form is represented by different matrices under different bases. Its algebraic form is , where must be an invertible matrix. This relation plays a foundational role in studying the properties and classification of quadratic forms.
Explore conditions, steps and evidence. Supplementary explanations are labeled separately from content shown in the video.
Write on the board , then write and .
The function values of a univariate quadratic function, with x as the independent variable.
The video does not specify the domain; formally, x is a real variable.
In the expression on the board, a, b, and c are coefficients of , x, and the constant term, respectively.
a,b,c
The coefficients in the general form of a univariate quadratic function; specifically, a is the quadratic coefficient, b is the linear coefficient, and c is the constant term.
The video does not specify the range; the image explicitly states the restriction .
Write on the board .
h,k
The translation parameters obtained by completing the square; h controls horizontal translation, and k controls vertical translation.
The video does not specify the range.
Write on the board .
,
The two independent variables of a bivariate quadratic function.
The video does not specify the domain; formally, they are real variables.
Write on the blackboard =ax_1^2+bx_1x_2+cx_2^2.
A binary quadratic function in variables and .
The video does not specify the domain.
The blackboard repeatedly displays multivariate quadratic expressions beginning with f=, used to determine whether they are quadratic forms.
f
The function name or left-hand side of the multivariate quadratic function in the example.
The video does not specify the domain.
In the example, appears, as in , , , and .
The third independent variable in the trivariate example.
The video does not specify the domain.
The blackboard repeatedly writes
f
a ternary quadratic form function
a ternary polynomial over the field of real numbers
The formula features , , and
, ,
the three variables of the quadratic form
real numbers
The blackboard writes and notes ()
A
the coefficient matrix corresponding to the quadratic form
an real matrix
The blackboard displays the row vector () and the column vector [; ; ]
x
a column vector composed of variables
three-dimensional real space
The screen repeatedly writes the quadratic form and its matrix representation starting with f.
f
quadratic form function
three-variable quadratic polynomial
The narrator says that quadratic functions were already studied in junior high school, with general form =ax^2+bx+c, and can be rewritten by completing the square as .
The blackboard sequentially displays =ax^2+bx+ and =.
The video first introduces the familiar quadratic function of one variable: its general form is , where the screen explicitly labels ; by completing the square, it can be rewritten in vertex form as . Here, h and k correspond to the coordinates of the vertex after translation.
a is not equal to 0
x is the single independent variable
The voiceover explains that translating the coordinate system moves the parabola's vertex to the origin, yielding , which is called a homogeneous quadratic function.
The blackboard displays , with the label "homogeneous quadratic function" beside it.
The video calls "homogeneous quadratic function." Its defining feature is that the expression contains no linear term or constant term—only the quadratic term remains; geometrically, this simplified form results from translating the original parabola so its vertex lands at the origin.
for quadratic functions of one variable
contains neither a linear term nor a constant term
The voiceover says we also commonly encounter multivariate quadratic functions, such as quadratic functions of two variables, which include not only terms like and , but also cross terms like .
The blackboard displays .
The video uses =ax_1^2+bx_1x_2+cx_2^2 to illustrate the multivariate case. Here, ax_1^2 and cx_2^2 are squared terms in single variables, while bx_1x_2 is a cross term involving two distinct variables. The voiceover emphasizes that the degree of the cross term is also two, since it is the product of two independent variables.
and are independent variables
The total degree of each term is 2
The voiceover states that if every term in the function expression has degree two, then the quadratic function is called homogeneous; such a homogeneous quadratic function is also called a quadratic form.
The blackboard writes: "If every term of a multivariate quadratic function has variable degree 2, then the quadratic function is called homogeneous—that is, a quadratic form."
The core definition given in the video is: a multivariate quadratic function is homogeneous—that is, a quadratic form—if every term has variable degree 2. In other words, a quadratic form may contain only squared terms and cross terms of exactly two variables; it must not contain linear terms, constant terms, cubic terms, or products of three or more variables.
Multivariate quadratic function
Each term has degree 2
Contains only squared terms and cross terms of exactly two variables
The voiceover uses "these three functions are all quadratic forms" and "the following four functions are not quadratic forms" to distinguish positive and negative examples.
The blackboard lists several expressions f= and circles with boxes the non-quadratic terms: , , , and +2.
The video's discriminant method checks the degree of each term one by one: if the degree of every term equals 2, then it is a quadratic form; if any term has degree not equal to 2, then it is not a quadratic form. In the example, , , , and +2 are circled respectively, corresponding to a cubic term, a triple cross-product term, a linear term, and a constant term.
Applicable to multivariate polynomial expressions
Requires checking the total degree of each term individually
The narrator says, "It is a quadratic form" and "We now have a basic understanding of quadratic forms."
The blackboard displays several expressions containing squared terms and cross terms.
The video introduces quadratic forms using concrete polynomials, emphasizing that they consist of squared terms and cross terms in the variables.
The objects under discussion are real-coefficient polynomials.
The variables are , ,
The narrator explains, "Split each cross term into two equal terms—that is, halve each cross term."
is split into , is split into , is split into
To write a quadratic form in symmetric matrix form, each cross term must be evenly split into two terms, which are then paired with corresponding squared terms.
Used to construct the symmetric matrix representation
The splitting must be equal.
Write on the board: () A [; ; ] =
Voiceover explains: "We have obtained the quadratic form in matrix representation, where the middle matrix is symmetric."
Express the quadratic form in a compact matrix form using the product of a row vector, a coefficient matrix, and a column vector.
A is a symmetric matrix
x is the column vector of variables
Voiceover emphasizes: "Only when expressed as the corresponding real symmetric matrix is the quadratic form representation unique."
Write on the board several non-symmetric matrix representations, then highlight the symmetric matrix with a box.
Although a quadratic form may admit multiple matrix representations, only the symmetric matrix representation is unique; thus, the symmetric matrix is conventionally adopted.
Restricted to the real number field
Require
The instructor says, "In fact, the real symmetric matrix corresponding to a quadratic form can be written down directly."
The screen displays , and writes several matrix product forms.
The bottom text states, "The coefficient of the quadratic term is placed at the a_{ii} position on the main diagonal, and the coefficient of the cross term is halved and placed at the corresponding a_{ij} and a_{ji} positions."
When writing a ternary quadratic form as with A required to be a real symmetric matrix, place the coefficients of the square terms directly on the main diagonal, and halve the coefficients of the cross terms before placing them symmetrically at positions a_{ij} and a_{ji}. The video demonstrates this rule using .
The quadratic form is a polynomial with real coefficients
The target matrix representation uses a real symmetric matrix
Coefficients of cross terms must be halved before being placed symmetrically
The screen simultaneously displays three different matrix product representations.
Text at the bottom reads: "The matrix representation of a quadratic form is not unique; there are infinitely many possible representations."
A single quadratic form can correspond to multiple matrix product representations; therefore, seeing an expression of the form does not imply that B is the unique matrix for that quadratic form.
The discussion concerns different matrix representations of the same quadratic form
Text at the bottom reads: "Only when expressed as a real symmetric matrix is the matrix representation of a quadratic form unique."
The instructor emphasizes: "Thus, we can study quadratic forms using methods for real symmetric matrices."
Among all matrix representations, the matrix representation corresponding to a quadratic form is unique only when the matrix is required to be real and symmetric.
Matrix representation restricted to real symmetric matrices
The voiceover states that since is a variable of the function, the exponents in the cross product are also quadratic, and thus it is also a homogeneous quadratic function.
The voiceover's phrase 'the exponents are also quadratic' is imprecise; the formula shown on screen directly verifies that the total degree of the term is 2.
In =ax_1^2+bx_1x_2+cx_2^2, the total degree of the cross term is 2, so this expression remains a homogeneous quadratic function where every term has degree 2.
and are independent variables
The term represents the product of two distinct independent variables
Applies to the cross term in the given binary expression.
The blackboard writes: 'A quadratic form may contain only squared terms and cross-product terms involving exactly two variables.'
The narrator says it is obvious that a quadratic form contains only squared terms of variables and cross terms of two variables.
A quadratic form may contain only squared terms and cross-product terms of two variables.
The expression is a multivariate quadratic function.
Each term has degree 2.
This holds for all expressions referred to as quadratic forms in the video.
Narrator: "Every quadratic form corresponds to a symmetric matrix... the corresponding matrix is a real symmetric matrix."
Blackboard: "Every real quadratic form corresponds to a real symmetric matrix."
Every real quadratic form corresponds uniquely to a real symmetric matrix.
The quadratic form has real coefficients.
Construct the matrix by halving the cross-term coefficients.
An arbitrary real quadratic form
Narrator: "Due to the arbitrariness of splitting, there are infinitely many ways for a matrix to represent a quadratic form."
Blackboard: "The matrix representation of a quadratic form is not unique; there can be infinitely many representations."
If symmetry of the matrix is not required, then the same quadratic form can have infinitely many matrix representations.
Allowing unequal splitting of cross terms
is not required.
There exist infinitely many representations.
The bottom text explicitly states: "Only when expressed as a real symmetric matrix is the matrix representation of a quadratic form unique."
The same quadratic form can be represented by infinitely many matrices; the matrix representation becomes unique only when restricted to real symmetric matrices.
We are discussing matrix representations of the same quadratic form.
The comparison involves general matrix representations and real symmetric matrix representations.
For all matrix representations of the same quadratic form,
The bottom text reads: "The rank of a quadratic form is defined as the rank of its corresponding real symmetric matrix."
The screen displays .
If a quadratic form f is expressed as with , then we define .
f has already been written in the form .
A is a real symmetric matrix.
For the quadratic form f and matrix A satisfying the above representation condition,
The instructor says: "Before computing the rank of a quadratic form, you must first check whether its corresponding matrix is real symmetric; if it is not real symmetric, you must first convert it into a real symmetric matrix before computing its rank."
The bottom text reads: "The matrix representation of a quadratic form is not unique; therefore, for a quadratic form expressed as , one must always check whether A is a real symmetric matrix."
Because the matrix representation of a quadratic form is not unique, one must first verify whether the matrix used is real symmetric before computing the rank of the quadratic form; if it is not, one must first convert it into a real symmetric matrix and then compute its rank.
The quadratic form is already written in the form
The goal is to find the rank of this quadratic form
For any quadratic form given in the form
Matrix congruence reflects the matrix representations of the same quadratic form under different bases, and like matrix similarity, it is also an equivalence relation on matrices.
Matrix congruence and matrix similarity are both special types of matrix equivalence relations.
Matrices A and B are congruent: B equals P transpose times A times P, where P is invertible.
For all matrices A and B.
The narrator says that by completing the square, we can write , and then translate the coordinate system so that the parabola's vertex moves to the origin, yielding = ax^2.
The blackboard sequentially displays = ax^2 + bx + , = , and = ax^2.
The video does not show the specific algebraic expansion steps from ax^2 + bx + c to .
Starting from the general form of a univariate quadratic function.
The starting formula given directly on the blackboard.
The narration explains that the expression can be rewritten in vertex form via completing the square.
The video states orally, 'written via completing the square', while simultaneously displaying the expression.
Then translate the vertex to the coordinate origin, eliminating the translation parameters h and k, yielding a form containing only the quadratic term.
The narration explicitly says, 'via coordinate translation... the vertex is translated to the coordinate origin, thus becoming '.
Using intuitive explanations of completing the square and coordinate translation, the video simplifies a general univariate quadratic function to , and from this introduces the concept of a 'homogeneous quadratic function'.
The upper half of the blackboard lists three expressions consisting entirely of quadratic terms; the lower half lists four expressions containing non-quadratic terms.
A white box sequentially highlights , , , and +2, ending with a red cross.
In the first example, all terms are , , , , , and , each having total degree 2.
Each term's total degree is checked one by one, conforming to the definition of a quadratic form given in the video.
The second example contains only pairwise cross terms, each still of degree 2.
Each term is the product of two independent variables.
The third example contains only square terms and also satisfies the condition that each term has degree 2.
Each term is the square of a single independent variable.
The fourth example contains , which has degree 3, and therefore is not a quadratic form.
It violates the definition requiring that every term have degree 2.
The fifth example contains , which is a product of three variables and has degree 3, and therefore is not a quadratic form.
It violates the requirement that only square terms and pairwise cross-product terms are allowed.
The sixth example contains , which is a linear term, and therefore is not a quadratic form.
It contains a term of degree 1.
The seventh example contains the constant term +2, which has degree 0, and therefore is not a quadratic form.
It contains a term whose degree is not 2.
The video confirms that the first three expressions are quadratic forms by checking the degree of each term, while the last four are not quadratic forms because they contain cubic terms, triple-variable cross products, linear terms, or constant terms.
Gradually rewrite the original quadratic form into grouped form, factored form, and matrix product form.
Voice-over simultaneously explains each algebraic step.
Write the original ternary quadratic form.
Given directly in the video.
Split each cross term in half and regroup by variable.
Voice-over states "split each cross term in half", and the chalkboard displays the splitting result.
Factor out the common variables , , and from each group.
Apply the distributive law in reverse.
Express as a row vector multiplied by a column vector.
Definition of the vector dot product.
Further decompose the column vector into the product of a coefficient matrix and a variable column vector.
Expansion of matrix multiplication.
Obtain the standard matrix representation of a quadratic form, where the middle matrix is symmetric.
Notation simplification and observation of symmetry.
By halving cross terms and using matrix multiplication, a real quadratic form can be uniquely expressed as , where A is a real symmetric matrix.
The instructor explains step by step: 'First, write the coefficients of the squared terms on the main diagonal; then halve the coefficient of the cross term—for example, —halving 4 to obtain 2, and placing these two 2s respectively in the first row second column and the second row first column.'
The screen displays the polynomial form of f and the final matrix [[1,2,1],[2,2,1],[1,1,3]].
Colored boxes and connecting lines match each polynomial term with its corresponding matrix entry position.
Starting from a given ternary quadratic form, first identify the square terms and the cross terms.
The video screen directly shows this polynomial.
Place the coefficients of the squared terms sequentially along the main diagonal.
The bottom rule states: 'The coefficient of the squared term is placed at position a_{ii} on the main diagonal.'
the cross term4x_1x_2 coefficient4 is halved to2, and placed symmetrically in row1 row number 2 column and row2 row number 1 column.
The instructor gives a clear example illustrating how to place this cross term.
the cross term2x_2x3 coefficient2 halved is1, and placed symmetrically in row2 column3 and row3 column2.
The instructor says, 'The other two cross terms follow the same rule.'
Cross term2x1x3 coefficient2 halved to 1 and symmetrically placed in row 1 column 3 and row 3 column 1.
The instructor says, 'The other two cross terms follow the same rule.'
Collecting the above elements yields a real symmetric matrix.
The final on-screen matrix satisfies .
The unique real symmetric matrix representation of this quadratic form is , where .
The instructor says, 'Next, let us examine two special cases: the first contains only squared terms and no cross terms; the second contains only cross terms and no squared terms.'
The screen displays two special quadratic forms along with their matrix representations.
Text at the bottom summarizes the matrix features of these two special cases.
First, consider the case containing only square terms.
The screen directly shows this quadratic form.
Since there are no cross terms, all off-diagonal entries are 0, yielding a diagonal matrix.
Text at the bottom states: "If there are no cross terms (only square terms), then the corresponding matrix is a simple diagonal matrix."
Next, consider the case containing only cross terms.
The screen directly shows this quadratic form.
Since there are no square terms, all main diagonal entries are 0; coefficients of cross terms are halved and placed symmetrically.
Text at the bottom states: "If there are no square terms (only cross terms), then all main diagonal entries of the corresponding matrix are 0."
When only square terms are present, the corresponding matrix is diagonal; when only cross terms are present, the main diagonal entries of the corresponding real symmetric matrix are all 0.
f equals x transpose times A times x, and x equals P times y, so f equals (P y) transpose times A times (P y), which is y transpose times P transpose times A times P times y, i.e., y transpose times B times y; therefore B equals P transpose times A times P.
According to the coordinate transformation formula, introduce the change-of-basis matrix P from basis C to basis D.
Relationship between basis change and coordinate transformation
Substitute x equals P y into the quadratic form's expression f equals x transpose A x under basis C.
Algebraic substitution
Use the reverse-order property of transpose: (A B) transpose equals B transpose A transpose, to expand (P y) transpose.
Matrix transpose operation laws
Compare f in its standard form under basis D, , to obtain the expression for matrix B.
Polynomial identity principle
The matrix representations of the same quadratic form under different bases satisfy the congruence relation .
The blackboard lists , , and .
The narrator says, "For example, these three functions are all quadratic forms."
Determine whether each of the following expressions is a quadratic form: ; ; .
All three expressions are written on the blackboard.
A quadratic form is defined as an expression in which every term has total degree 2.
Check the degree of each term individually and determine whether the expression is a quadratic form.
Each term in the first expression is either the square of a single variable or the cross product of two variables.
The total degree of each of these terms is 2.
The second expression contains only pairwise cross terms.
The total degree of each term is 2.
The third expression contains only squared terms.
The total degree of each term is 2.
All three expressions are quadratic forms.
Consistent with the video narration: "These three functions are all quadratic forms."
The blackboard lists , , , .
A white box sequentially highlights , , , and +2, ending with a red cross.
The narration states that the following four functions are not quadratic forms and cites cubic terms, triple cross-product terms, linear terms, and constant terms.
Determine whether the following expressions are quadratic forms: ; ; ; .
All four expressions have been written on the blackboard.
A quadratic form requires every term to have degree 2.
Identify the term violating the quadratic form condition and make a judgment.
The first expression contains , which is a cubic term.
Its degree is 3, not equal to 2.
The second expression contains , a term where three variables are multiplied together.
The total degree is 3, and it does not satisfy 'a cross-product term involving any two variables'.
The third expression contains , a linear term.
Its degree is 1, which is not equal to 2.
The fourth expression contains the constant term +2.
The degree of the constant term is 0, which is not equal to 2.
None of the four expressions is a quadratic form.
This matches the video narration 'None of the following four functions is a quadratic form' and the red cross marks on screen.
The blackboard derivation continuously revolves around .
Write in matrix form.
f is a ternary real quadratic form.
Contains squared terms and cross terms.
Find the corresponding symmetric matrix A such that .
Split the cross terms evenly.
To ensure matrix symmetry, halve the coefficients.
Read off the coefficients of each variable combination to form the matrix.
The diagonal entries are the coefficients of the squared terms; the off-diagonal entries are the halved coefficients of the cross terms.
f equals x transpose times a three-by-three matrix times x, where the matrix has first row one, two, one; second row two, two, one; third row one, one, three.
Expanding recovers the original polynomial, and A equals A transpose.
The screen displays and the matrix [[1,2,1],[2,2,1],[1,1,3]].
The instructor explains term by term how the squared terms and cross terms are placed into the matrix.
Colored boxes and connecting lines indicate the correspondence between terms and matrix entries.
Write in the form , making A a real symmetric matrix.
f equals the square of x one plus two times the square of x two plus three times the square of x three plus four times x one times x two plus two times x two times x three plus two times x one times x three.
The matrix representation must be real and symmetric.
Write the corresponding real symmetric matrix A.
Coefficients of squared terms go on the main diagonal.
The writing rule shown in the video.
coefficient 4 halved to 2 and symmetrically placed.
The instructor gives an explicit example.
coefficient2 halved is1 placed symmetrically.
The instructor explains that the remaining cross terms follow similarly.
2x1x3 coefficient2 halved is1 placed symmetrically.
The instructor explains that the remaining cross terms follow similarly.
Summarizing yields the matrix.
The screen finally displays this matrix.
A is a three-by-three matrix whose first row entries are one, two, one; second row entries are two, two, one; third row entries are one, one, three.
The screen simultaneously shows the polynomial and its matrix product form, with the matrix satisfying .
The screen displays and its matrix representation.
The bottom text states that when only square terms are present, the corresponding matrix is diagonal.
Express in real symmetric matrix form.
f equals the square of x₁ plus two times the square of x₂ plus three times the square of x₃
without cross terms
write the corresponding matrix and observe its structure
place the coefficients of the squared terms on the main diagonal, and fill all other positions with 0.
video rules and bottom text explanation.
A is a three-by-three matrix whose first row elements are one, zero, zero; second row elements are zero, two, zero; third row elements are zero, zero, three
the video directly displays this matrix product form, and it is a diagonal matrix.
the video displays f equals plus plus and its matrix representation.
the bottom text explanation states that when only cross terms are present, all main diagonal entries are 0.
express f equals plus plus in real symmetric matrix form.
f equals four times x₁ times x₂ plus two times x₂ times x₃ plus two times x₁ times x₃
without square terms
write the corresponding matrix and examine its main diagonal
there are no square terms, so all entries on the main diagonal are 0; coefficients of cross terms are halved and placed symmetrically.
video rules and bottom text explanation.
A is a three-by-three matrix whose first row entries are zero, two, one; second row entries are two, zero, one; third row entries are one, one, zero
the video directly shows this matrix product form, and all main diagonal entries are 0.
At the beginning, 'Popular Matrix' and 'Professor Xiao Zhiche Gao Shan' appear, followed by a green blackboard title page reading 'Lecture 44: The Concept of Quadratic Forms'.
'Popular Matrix'
'Professor Xiao Zhiche Gao Shan'
Blackboard title: "Lecture 44: Concept of Quadratic Forms"
Transition from light-background opening sequence to green blackboard title page
No mathematical content
This segment serves as course introduction, contains no mathematical reasoning, and only states the lecture number and topic.
First appears on the blackboard: =ax^2+bx+; then below it appears =; finally appears =ax^2 together with the label "homogeneous quadratic function".
equals a times x squared plus b times x plus c, where a is not equal to 0
equals a times the quantity x minus h squared, plus k
equals a times x squared
Text: "homogeneous quadratic function"
Formulas appear line by line from top to bottom
Transition from general form to vertex form, then to the form containing only the quadratic term
Always centered on the quadratic function in one variable.
The visual order matches the lecture sequence: first reviewing the general form, then presenting the completed-square form, and finally abstracting the homogeneous quadratic function .
A new expression appears below the blackboard: , followed by two yellow text lines about the definition of quadratic forms.
f acting on x₁ and x₂ equals a times the square of x₁ plus b times x₁ times x₂ plus c times the square of x₂.
Definition text: "If every term of a multivariate quadratic function has total variable degree 2, then this quadratic function is called homogeneous—that is, a quadratic form."
Text: "A quadratic form may contain only squared terms and cross-product terms involving exactly two variables."
The independent variables expand from x to , .
The cross term is added.
Definition text appears line by line.
The theme remains that each term has degree 2.
The screen generalizes the univariate case to the multivariate case and explicitly states the criterion for determining a quadratic form.
The upper half of the blackboard lists three positive examples; the lower half lists four negative examples; white boxes sequentially highlight , , , and +2, followed finally by a red cross.
Three quadratic form expressions
Four non-quadratic form expressions
White highlight box
Red cross
Positive examples are shown first, then negative examples
White boxes mark, one by one, the non-quadratic terms in the negative examples
A red cross finally summarizes the negative judgment
All examples begin with f=, facilitating term-by-term comparison of degrees
The animation grounds the abstract definition in concrete expressions: all positive examples satisfy the condition that each term has degree 2; negative examples violate this condition, with violating terms highlighted.
Multiple quadratic forms are listed on the green chalkboard, with some terms highlighted in white boxes and red X marks on the right.
Multiple f expressions.
White box.
Red X mark.
Different expressions displayed side by side.
Some terms are highlighted with boxes.
All expressions are in forms related to quadratic forms.
By comparison, the typical structure of a quadratic form is shown, and some notations are implied to violate the subsequent symmetric matrix requirement.
The formula appears line by line, from polynomial to grouping, factoring, and matrix product.
Polynomial.
Grouping with parentheses.
Row vector.
Column vector.
matrix.
Cross terms are separated.
The expression is reorganized.
Introduce matrix notation
The equation always holds
The set of variables remains unchanged
Visualize the algebraic identity transformation process to help understand how to transition from polynomial form to matrix form.
The blackboard lists several different matrices, finally highlighting the symmetric matrix with a white box
Multiple matrices
White highlight box
Show various matrix representations
Finally focus on the unique symmetric representation
All matrices correspond to the same quadratic form
Explain that the general matrix representation is not unique, whereas the symmetric matrix representation is unique.
a green blackboard simultaneously displays a quadratic form and three matrix product representations, with two yellow textual conclusions at the bottom.
quadratic form f
three 3 by 3 matrices
row vector (x subscript 1, x subscript 2, x subscript 3)
column vector (x subscript 1, x subscript 2, x subscript 3) transpose
Two concluding statements at the bottom
Static on-screen display showing multiple representations side by side for comparison
All three correspond to the same quadratic form f
The side-by-side display illustrates that the matrix representation of a quadratic form is not unique, while the real symmetric representation is specially highlighted with a box.
Other matrices fade out, leaving only the real symmetric matrix representation with a white border.
Each term of the polynomial is marked with a colored square and connected by lines to its corresponding element in the matrix.
Terms of the polynomial
Real symmetric matrix [[1,2,1],[2,2,1],[1,1,3]]
Colored connecting lines
White outer border
First highlight the final real symmetric matrix
Then connect the quadratic terms and cross terms separately to their corresponding matrix positions.
The matrix always maintains a symmetric structure.
The visual process concretizes the abstract rule: quadratic terms correspond to the main diagonal, and cross terms—halved—correspond to pairs of symmetric positions.
The screen switches to two special quadratic forms and their matrix representations.
Text at the bottom summarizes the matrix features of these two special cases.
f equals the square of x₁ plus two times the square of x₂ plus three times the square of x₃.
f equals four times x₁ times x₂ plus two times x₂ times x₃ plus two times x₁ times x₃.
Diagonal matrix.
A real symmetric matrix with zeros on the main diagonal.
Switching from a general example to two special examples.
Giving the corresponding matrices separately.
still using the same writing rule for real symmetric matrices
illustrating the marked structural differences between matrix forms containing only square terms and only cross terms, by side-by-side comparison
the screen returns to the general example, displaying 〔〕, 〔〕, and 〔〕
hints about the definition of rank and checking real symmetry appear sequentially at the bottom
x transpose times A times x
condition 〔〕
equality 〔〕
hint text at the bottom
first presenting the matrix representation and the symmetry condition
then giving the defining formula for rank
finally adding that symmetry must be verified before computing rank
The object under discussion remains the same quadratic form f.
The screen converts 'the rank of a quadratic form' into 'the rank of its corresponding real symmetric matrix' and stresses that symmetry is a prerequisite for this definition to hold.
A table appears on the blackboard listing correspondences among quadratic forms, basis vector sets, and coordinates.
Table
Quadratic form f
Basis vector sets C and D
Coordinates x and y
The table displays row by row the expressions of quadratic form f under different bases and their corresponding coordinate variables.
The quadratic form f itself remains unchanged; only the basis vectors and coordinate variables change.
Intuitively demonstrates that the matrix representation and coordinate variables of the same quadratic form change depending on the chosen basis vectors.
The voiceover specifically explains that although is a cross term, it is still quadratic because it involves the product of two variables.
When seeing cross terms like , one easily mistakes them for non-quadratic terms, unlike or .
The video explicitly states that is the product of two independent variables, and its total degree remains 2, so it may appear in a quadratic form.
In the counterexamples, although most terms are square terms or cross terms, the presence of any of , , , or +2 causes the expression to be ruled out as a quadratic form.
One easily misinterprets "containing quadratic terms" as "the entire expression being a quadratic form".
The video uses counterexamples to clarify that a quadratic form requires every term to have degree exactly 2; mixing in linear terms, constant terms, cubic terms, or triple products disqualifies it as a quadratic form.
The narrator warns, "If we do not apply halving... we can still obtain a matrix, but it will not be a real symmetric matrix."
The blackboard displays several non-symmetric matrices.
The belief that as long as a quadratic form can be written as , the matrix B may be chosen arbitrarily and without distinction.
The video points out that non-symmetric representations are not unique, whereas the symmetric matrix representation is both unique and standard.
The instructor reminds viewers, 'A quadratic form has infinitely many matrix representations; therefore, before computing the rank of a quadratic form, one must always verify whether the corresponding matrix is a real symmetric matrix.'
The bottom text reads: 'The matrix representation of a quadratic form is not unique; thus, for a quadratic form expressed as , one must always check whether A is a real symmetric matrix.'
Assuming that, upon seeing , B must be 'that' matrix of the quadratic form and can be used directly to compute the rank.
The matrix representation of a quadratic form is not unique; only the real symmetric matrix representation is unique. Before computing the rank of a quadratic form, one must first confirm that the matrix is real symmetric; otherwise, it must first be converted into a real symmetric matrix.
The instructor explicitly explains, 'Halve the coefficients of cross terms—for example, here 4, 2, and 2 become two sets: 2, 1, and 1.'
The screen maps to the two entries 2 in the matrix.
Placing the coefficient of a cross term unchanged into a single matrix entry, rather than halving it and placing the result symmetrically into two entries.
In the real symmetric matrix representation, the coefficient of the cross term must be halved and placed respectively at the symmetric positions a_{ij} and a_{ji}.
The matrix representation of a quadratic form is not unique; thus, for a quadratic form expressed as , one must verify whether A is a real symmetric matrix.
The belief that a quadratic form corresponds to only one unique matrix.
The matrix representation of a quadratic form is not unique; one must verify whether its corresponding matrix is a real symmetric matrix before proceeding to discuss rank and congruence transformations.
The voiceover first introduces the general form = ax^2 + bx + c, then explains completing the square and translation to obtain = ax^2.
The three forms on the blackboard are arranged in order.
= ax^2 is a simplified special case of a univariate quadratic function obtained by completing the square and translation; the video names it a homogeneous quadratic function.
The voiceover states that for a univariate quadratic function, = ax^2 is a homogeneous function, but one also commonly encounters multivariate quadratic functions, and gives a bivariate example.
The blackboard transitions from = ax^2 to = ax_1^2 + bx_1x_2 + cx_2^2.
The binary example generalizes the univariate homogeneous quadratic function to the multivariable case and introduces cross terms .
First, state the definition of a quadratic form, then list positive and negative examples and circle the violating terms.
The judgment method applies the definition directly: check term by term whether all degrees equal 2.
White boxes and red crosses apply the method to concrete expressions.
The counterexamples demonstrate how to exclude non-quadratic forms using the term-by-term degree-checking method.
The narrator directly links "halving" with "symmetric matrix".
Halving the cross terms is a necessary preliminary step for constructing a symmetric matrix representation.
First present , then discuss the uniqueness of A.
The matrix representation includes further requirements regarding symmetry and uniqueness.
First state the writing rule, then emphasize that the real symmetric matrix representation is unique.
Direct application of the writing rule yields the unique real symmetric matrix representation.
The bottom text directly links 'non-unique representation' with 'must check whether A is real symmetric'.
Because the matrix representation is not unique, we must first verify that the matrix used is real symmetric before defining or using the rank of a quadratic form.
The general rule is stated first, followed by a special case containing only square terms.
Containing only square terms is a special case of the general writing rule: when there are no cross terms, all off-diagonal entries equal 0, yielding a diagonal matrix.
The general rule is stated first, followed by a special case containing only cross terms.
Containing only cross terms is also a special case of the general writing rule: when there are no square terms, all main diagonal entries equal 0.
The screen first displays with (), then .
The rank of a quadratic form can be defined as only because of the uniqueness of the real symmetric matrix representation.
Like similarity, congruence is also an equivalence relation between matrices.
This video draws an analogy between matrix congruence and matrix similarity, pointing out that both are special types of matrix equivalence relations—congruence corresponds to representations of quadratic forms under different bases, while similarity corresponds to representations of linear transformations under different bases.
The blackboard writes the definition of a quadratic form directly.
The voiceover explains that represents cross multiplication and is also of degree two.
White boxes circle the violating terms: , , , and +2.
The counterexamples contain, respectively, a cubic term, a product of three terms, a linear term, and a constant term.
The narrator says that by completing the square and translating coordinates, the vertex can be moved to the origin, yielding .
The narrator explains the purpose of halving.
The blackboard explicitly writes the contrast between "not unique" and "unique".
The instructor explains step by step how to write the real symmetric matrix from a quadratic form.
The presenter explains halving the coefficient of the cross term and placing it symmetrically.
The bottom text states that the matrix representation of a quadratic form is not unique, but the real symmetric representation is unique.
The screen shows an example containing only square terms and its diagonal matrix.
The screen shows an example containing only cross terms and its matrix with zeros on the main diagonal.
Covered · Opening sequence and title screen; no mathematical content.
Covered · Review the standard and completed-square forms of univariate quadratic functions.
Covered · Introduce the univariate homogeneous quadratic function .
Covered · Extend to bivariate quadratic functions and explain square terms and cross terms.
Covered · State the definition of a quadratic form and the types of terms it may contain.
Covered · List three positive examples of quadratic forms.
Covered · List four counterexamples that are not quadratic forms, with highlighted boxes and red crosses explaining why.
Covered · Introduce the concept of quadratic forms and display multiple expressions.
Covered · Demonstrate halving cross terms and grouping in detail.
Covered · Derive into the form .
Covered · Emphasize the correspondence with real symmetric matrices.
Covered · Compare multiple matrix representations and identify the uniqueness condition.
Covered · Demonstrate multiple matrix representations of the same quadratic form, and point out that only the real symmetric representation is unique.
Covered · Use a general example to illustrate step by step how to write the real symmetric matrix from a quadratic form.
Covered · Present two types of special quadratic forms and summarize their matrix structures.
Covered · Define the rank of a quadratic form, and emphasize that we must first confirm the matrix is real symmetric before computing its rank.
Covered · Explains the definition of the rank of a quadratic form and cautions about the non-uniqueness of its matrix representation.
Covered · Display the quadratic form's expressions and coordinate variables under different bases via a table.
Covered · Derive the congruence relation for quadratic form matrices under basis change: .
Covered · Define congruence transformation and explain the matrix relationship of a quadratic form under different bases.
Covered · Explain that congruence is an equivalence relation on matrices, and draw an analogy with similarity.
Covered · Introduce the diagonalization goal: find an invertible matrix P such that A is congruent to a diagonal matrix U, thereby reducing the quadratic form to a special form containing only square terms.
Covered · Explaining matrix congruence and quadratic form diagonalization, and raising two related questions.
Covered · Video ends, displaying the THANK YOU screen.
Reviewed subject paths
Continue with related videos in the same subject.
